Oligodendrocyte-lineage cell exocytosis and L-type prostaglandin D synthase promote oligodendrocyte development and myelination

In the developing central nervous system, oligodendrocyte precursor cells (OPCs) differentiate into oligodendrocytes, which form myelin around axons. Oligodendrocytes and myelin are essential for the function of the central nervous system, as evidenced by the severe neurological symptoms that arise in demyelinating diseases such as multiple sclerosis and leukodystrophy. Although many cell-intrinsic mechanisms that regulate oligodendrocyte development and myelination have been reported, it remains unclear whether interactions among oligodendrocyte-lineage cells (OPCs and oligodendrocytes) affect oligodendrocyte development and myelination. Here, we show that blocking vesicle-associated membrane protein (VAMP) 1/2/3-dependent exocytosis from oligodendrocyte-lineage cells impairs oligodendrocyte development, myelination, and motor behavior in mice. Adding oligodendrocyte-lineage cell-secreted molecules to secretion-deficient OPC cultures partially restores the morphological maturation of oligodendrocytes. Moreover, we identified L-type prostaglandin D synthase as an oligodendrocyte-lineage cell-secreted protein that promotes oligodendrocyte development and myelination in vivo. These findings reveal a novel autocrine/paracrine loop model for the regulation of oligodendrocyte and myelin development.


Introduction
In the developing central nervous system (CNS), oligodendrocyte precursor cells (OPCs) differentiate into oligodendrocytes (Bergles and Richardson, 2016;Hill et al., 2014;Kang et al., 2010), which form myelin sheaths around axons. Myelin is essential for the propagation of action potentials and for the metabolism and health of axons (Fünfschilling et al., 2012;Larson et al., 2018;Mukherjee et al., 2020;Saab et al., 2016;Schirmer et al., 2018;Simons and Nave, 2016). When oligodendrocytes and myelin are damaged in demyelinating diseases such as multiple sclerosis (MS) and leukodystrophy, sensory, motor, and cognitive deficits can ensue (Gruchot et al., 2019;Lubetzki et al., 2020;Stadelmann et al., 2019). In a broader range of neurological disorders involving neuronal loss, such as brain/spinal cord injury and stroke, the growth and myelination of new axons are necessary for neural repair (Wang et al., 2020). Thus, understanding oligodendrocyte development and myelination is critical for developing treatments for a broad range of neurological disorders.
To determine whether cell-cell interactions within the oligodendrocyte lineage regulate oligodendrocyte development, we blocked VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells in vivo and found impairment in oligodendrocyte development, myelination, and motor behavior in mice. Similarly, exocytosis-deficient OPCs exhibited impaired development in vitro. Adding oligodendrocyte-lineage cell-secreted molecules promoted oligodendrocyte development. These results suggest that an autocrine/paracrine loop promotes oligodendrocyte development and myelination. We assessed L-PGDS as a candidate autocrine/paracrine signal and further discovered that oligodendrocyte development and myelination were impaired in L-PGDS-knockout mice. Moreover, overexpression of the gene encoding L-PGDS partially restored the myelination defect of exocytosisdeficient mice. These results reveal a new autocrine/paracrine loop model for the regulation of oligodendrocyte development in which VAMP1/2/3-dependent exocytosis from oligodendrocyte-linage cells and secreted L-PGDS promote oligodendrocyte development and myelination.

Expression of botulinum toxin B in oligodendrocyte-lineage cells in vivo
If oligodendrocyte-lineage cells use autocrine/paracrine mechanisms to promote development and myelination, one would predict that (1) blocking secretion from oligodendrocyte-lineage cells would impair oligodendrocyte development and myelination and, in turn, that (2) adding oligodendrocytelineage cell-secreted molecules might promote oligodendrocyte development. Membrane fusion relies on soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors (SNARE) family proteins located on vesicles (v-SNAREs) and target membranes (t-SNAREs). The binding of v-SNAREs and t-SNARES form intertwined α-helical bundles that generate force for membrane fusion (Pobbati et al., 2006). VAMP1/2/3 are v-SNAREs that drive the fusion of vesicles with the plasma membrane to mediate exocytosis (Chen and Scheller, 2001). We found that oligodendrocyte-lineage cells express high levels of VAMP2 and VAMP3 and low levels of VAMP1 in vivo (Figure 1A-C; Zhang et al., 2014;Zhang et al., 2016), consistent with previous reports in vitro (Feldmann et al., 2009;Feldmann et al., 2011;Madison et al., 1999). We found that VAMP2 + and VAMP3 + puncta are distributed throughout cultured OPCs, including.
To block VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells, we crossed Pdgfra-CreER transgenic mice, which express Cre recombinase in OPCs (PDGFRα + Olig2 + ; Kang et al., 2010), with loxP-stop-loxP-botulinum toxin B light chain-IRES-green fluorescent protein (GFP) (inducible botulinum toxin B, or ibot) transgenic mice (Slezak et al., 2012), allowing expression of botulinum toxin B-light chain in OPCs and their progeny. The light chain contains the catalytically active domain of the toxin but lacks the heavy chain, which allows cell entry (Montal, 2010), thus confining toxin expression to the targeted cell type. Therefore, the ibot transgenic mice allow for the inhibition of VAMP1/2/3-dependent exocytosis in a cell-type-specific and temporally controlled manner (Slezak et al., 2012).
In our study, we used double-transgenic mice hemizygous for both Cre and ibot and referred to them as the PD:ibot mice thereafter. To validate our model and test its recombination efficiency, we injected 0.1 mg of 4-hydroxytamoxifen in each PD:ibot mouse daily for 2 days between postnatal days 2-4 (P2-4) and examined GFP expression at P8 and P30. We assessed whether GFP expression is restricted to oligodendrocyte-lineage cells (specificity) and what proportion of oligodendrocyte-lineage cells express GFP (efficiency/coverage). At P8, when the vast majority of oligodendrocyte-lineage cells are undifferentiated OPCs, we detected specific expression of GFP in oligodendrocyte-lineage cells ( Figure 1D-F). GFP was efficiently expressed by oligodendrocyte-lineage cells throughout the brain, including the cerebral cortex grey matter, corpus callosum, and striatum ( Figure 1G). At P30, when substantial numbers of OPCs have differentiated (PDGFRα -Olig2 + ), we observed a similarly high specificity of GFP expression in oligodendrocyte-lineage cells (Figure 1-figure supplement 2). These observations are consistent with previous reports on the specificity and efficiency of the Pdgfra-CreER transgenic line (Kang et al., 2010). As controls, we used wildtype mice, mice with only the Cre transgene or only the ibot transgene subjected to the same tamoxifen injection scheme. In all control conditions, we detected very little GFP expression ( Figure 1-figure supplement 3).
To directly assess the expression of botulinum toxin B-light chain and the cleavage of the VAMP proteins in oligodendrocyte-lineage cells from PD:ibot mice, we purified OPCs from PD:ibot and control mice by immunopanning and allowed them to differentiate into oligodendrocytes in culture. We performed western blot analysis of the cultures and detected botulinum toxin B-light chain in PD:ibot but not in control cells ( Figure 1H). Furthermore, the levels of full-length VAMP2 and VAMP3 proteins were lower in PD:ibot cells compared with control cells ( Figure 1I, J, L and M). Based on these observations, we conclude that the botulinum toxin-GFP transgene is specifically and efficiently expressed by oligodendrocyte-lineage cells in PD:ibot mice.
To determine the stage(s) of oligodendrocyte development affected by botulinum toxin, we performed RNAscope in situ hybridization in vivo using probes for Enpp6, a marker for pre-myelinating oligodendrocytes, and Mbp, a marker for oligodendrocytes. Interestingly, both markers showed a significant reduction in PD:ibot mice compared with controls, demonstrating that the oligodendrocyte development defect in PD:ibot mice manifest as early as the pre-myelinating stage ( Figure 2I, J and K).
We next examined myelin development in PD:ibot mice and found that immunofluorescence of myelin basic protein (MBP), one of the main components of CNS myelin, is reduced in PD:ibot mice ( Figure 3A-H). Moreover, many MBP + ibot-GFPexpressing cells exhibit round cell morphology whereas MBP + GFPcontrol cells form elongated myelin internodes along axon tracks ( Figure 3I). Transmission electron microscopy allows for the assessment of myelin structure at a high resolution. Thus, to further examine myelination in PD:ibot mice, we performed transmission electron microscopy imaging and found a reduction in the percentage of myelinated axons (Fig. 3 J, L) and reduced myelin thickness in PD:ibot mice (g-ratio: axon diameter divided by the diameter of axon +myelin; Figure 3K and M). The density and diameter of axons did not differ between PD:ibot and control mice ( Figure 3N and O).
To determine whether botulinum toxin-B-expressing cells contribute to the population of surviving differentiated oligodendrocytes, we examined the overlap between GFP + botulinum-expressing cells and differentiated oligodendrocytes (Olig2 + PDGFRαcells and CC1 + cells) and found that botulinumexpressing cells can survive and become differentiated oligodendrocytes at P8 and P30 (Figure 3figure supplement 2A). We next compared the ratio of CC1 + cells to Olig2 + cells in GFP + cells in PD:ibot mice, GFPcells in PD:ibot mice, and all cells (GFP -) in control cells and did not observe statistically significant difference between any of the groups (Figure 3-figure supplement 2B). To examine whether OPC proliferation is affected in PD:ibot mice, we quantified the percentage of Ki-67 + proliferating cells among PDGFRα + OPCs and found an increase in OPC proliferation (  Source data 1. It contains original data points for Figure 1A        To investigate whether the expression of botulinum toxin B-light chain affects oligodendrocyte development and myelination in non-cell-type-specific manners, we blocked exocytosis from astrocytes or endothelial cells by crossing ibot transgenic mouse with Gfap-Cre (line 77.6) and Tek-Cre strains, respectively. We found astrocyte-and endothelial cell-specific expression of ibot-GFP in these mice but did not detect any obvious changes in oligodendrocyte density or myelin proteins. These observations suggest that botulinum toxin B-light chain peptides have specific effects on the targeted cell types. However, we could not rule out the possibility that embryonic and/or early postnatal compensation may mask the effect when a constitutive Cre is used.
To examine the functional consequences of blocking VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells, we assessed the motor behavior of PD:ibot and littermate control mice using the rotarod test. We placed mice on a gradually accelerating rotarod and recorded the time each mouse stayed on the rotarod. We found that PD:ibot mice stayed on the rotarod for significantly shorter amounts of time than littermate control mice on all 3 days of testing ( Figure 3P-S). Therefore, blocking VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells led to deficits in neural circuit function.

PD:ibot mice exhibit changes in the transcriptomes of OPCs and oligodendrocytes
We next aimed to uncover the molecular changes in OPCs and oligodendrocytes in PD:ibot mice. We performed immunopanning to purify OPCs and oligodendrocytes from the brains of P17 PD:ibot and littermate control mice and performed RNA-sequencing (RNA-seq). We detected broad and robust gene expression changes in oligodendrocytes, and moderate changes in OPCs ( Figure 4A and B, Supplementary files 1 and 2), demonstrating that VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells is critical for establishing and/or maintaining the normal molecular attributes of oligodendrocytes and OPCs. Notably, the expression of signature genes of differentiated oligodendrocytes such as Plp1, Mbp, Aspa, and Mobp was significantly reduced in oligodendrocytes purified from PD:ibot mice compared with controls ( Figure 4E-H, Supplementary files 1 and 2). This result was not secondary to a reduction in oligodendrocyte density, as we loaded a similar amount of cDNA libraries from PD:ibot and control oligodendrocytes for sequencing, and processed all sequencing data with the same pipeline. Therefore, VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells is critical for the expression of mature oligodendrocyte genes. The expression of genes encoding oligodendrocyte-lineage cell marker proteins used for immunohistochemistry, Qk (encoding the Qk protein recognized by the CC1 antibody), Olig2, and PDGFRα did not change in PD:ibot mice (Figure 4-figure supplement 1). Therefore, our oligodendrocyte count is not confounded by changes in marker gene expression. We next performed gene ontology (GO) analysis to reveal the molecular pathways and cellular processes altered in each type of glial cell in PD:ibot mice ( Figure 4I-L, Supplementary file 3). Genes associated with filopodium assembly, calcium ion transport, and plasma membrane raft assembly pathways were increased, whereas genes associated with lipid biosynthetic process, axon ensheathment, and myelination pathways were reduced in oligodendrocytes in PD:ibot mice. Genes associated with the trans-synaptic signaling, chemical synaptic transmission, and GPCR signaling pathways were increased in OPCs in PD:ibot mice. To assess whether oligodendrocyte-lineage cell exocytosis affects other glial cell types, such as astrocytes and microglia, we also purified these cells by immunopanning and performed RNA-seq. We observed moderate at P30. (C, F) Olig2 + oligodendrocyte-lineage cells and PDGFRα + OPCs in the cerebral cortex of PDibot and control mice at P8 (C) and P30 (F). Scale bars: 20 μm. (D, E, G, H) Quantification of the densities of Olig2 + and PDGFRα + cells in the cerebral cortex of PD:ibot and control mice at P8 and P30. N=5 mice per group at P8. N=4 mice per group at P30. Paired two-tailed T-test. Olig2 + cells/mm 2 : 535.6±73.6 in control and 444.4±82.9 in PD:ibot, p=0.040 at P8; 309.9±21.7 in control and 253.8±12.2 in PD:ibot, p=0.048 at P30. PDGFRα + cells/mm 2 : 358.4±22.8 in control and 350.9±34.0 in PD:ibot, p=0.74 at P8; 157.8±2.6 in control and 155.6±8.5 in PD:ibot, p=0.85 at P30. (I) Enpp6 + cells and Mbp + cells in the brains of PDibot and control mice at p15. Scale bar: 200 μm. (J, K) Quantification of the density of Enpp6 + cells and Mbp + cells in the cerebral cortex of PD:ibot and control mice at P15. N=4 mice per group. All control mice are ibot only (black dots). Paired two-tailed T-test. Enpp6 + cells/mm 2 : 91.4±14.1 in control and 17.5±6.2 in PD:ibot, p=0.015; Mbp + cells/mm 2 : 66.3±8.9 in control and 11.3±4.9 in PD:ibot, p=0.0057.
The online version of this article includes the following source data for figure 2: Source data 1. It contains original data points for Figure 2B, D, E, G, H, J and K.  Blocking VAMP1/2/3-dependent exocytosis from oligodendrocytelineage cells impairs oligodendrocyte development in vitro VAMP1/2/3-dependent exocytosis from oligodendrocyte-lineage cells may directly affect oligodendrocyte development or change the attributes of other cell types, and, in turn, indirectly affect oligodendrocytes. For example, OPC-secreted molecules may affect axonal growth, and subsequently, axonal signals may affect oligodendrocytes indirectly. Therefore, we next employed purified OPC and oligodendrocyte cultures to determine whether exocytosis has direct roles in oligodendrocyte-lineage cells in the absence of other cell types.
We performed immunopanning to purify OPCs from P7 PD:ibot and control mice injected with 4-hydroxytamoxifen as described above. We cultured the OPCs for two days in the proliferation medium and then switched to the differentiation medium and cultured them for another seven days. To assess oligodendrocyte differentiation and maturation, we assessed the levels of MBP protein, a marker for differentiated oligodendrocytes, and detected lower MBP levels in PD:ibot cells compared with controls ( Figure 5B and C). Similarly, Mbp mRNA levels were also lower as determined by quantitative real-time PCR ( Figure 5D).
Additionally, we assessed the morphological maturation of oligodendrocytes in vitro ( Figure 5E-H). OPCs are initially bipolar, and as they differentiate, they grow a few branches to become star-like. The cells next grow more branches to become arborized and then extend myelin-sheath-like flat membranous structures, acquiring a 'lamellar' morphology ( Figure 5A; Zuchero et al., 2015) (also referred to as a 'fried egg' or 'pancake' morphology). We used the CellMask dye to analyze the morphological maturation of oligodendrocytes. At day 3 of differentiation, we found that a larger proportion of PD:ibot cells than control cells are at the early 'star', stage whereas a smaller proportion of PD:ibot cells than control cells have proceeded to the late 'lamellar' stage ( Figure 5F). At day 7 of differentiation, more PD:ibot cells have proceeded from the 'star' to the 'arborized' stage compared with day 3, but the percentage of cells that have proceeded to the late 'lamellar' stage remains lower in PD:ibot p=0.024 at P30. (I) The morphology of ibot-GFP + cells and GFPcontrol cells labeled by MBP immunofluorescence. A region in the cerebral cortex from a P8 PD:ibot mouse is shown. The arrowheads point to ibot-GFP + cells and the arrows point to GFPcontrol cells. Scale bar: 50 μm. (J) Transmission electron microscopy images of the corpus callosum at P17. Scale bar: 1 μm. (K) g-ratio as a function of axon diameter in the corpus callosum at P17. N=3 mice per group. 112 myelinated axons from control and 97 myelinated axons from PD:ibot were analyzed. (L, M) Quantification of the percentage of myelinated axons and g-ratio (axon diameter divided by the diameter of myelin +axon) from the transmission electron microscopy images of the corpus callosum at P17. N=3 mice per group. All control mice at P17 are ibot only (black dots). Paired two-tailed T-test. Myelinated axons %: 7.6±0.7 in control and 2.0±0.4 in PD:ibot, p=0.0048. g-ratio: 0.84±0.009 in control and 0.92±0.0009 in PD:ibot; p=0.012. (N, O) Quantification of axon density and axon diameter from the transmission electron microscopy images of the corpus callosum at P17. N=4 mice per group. Paired two-tailed T-test. Axon number/µm 2 : 6.2±1.5 in control and 3.9±0.47 in PD:ibot, p=0.19. Axon diameter: 309.1±44.3 nm in control and 349.8±16.0 nm in PD:ibot; p=0.51. (P) Latency to fall from an accelerating rotarod (seconds). Each mouse was tested three times per day for 3 consecutive days. The average latency to fall of the three trials of each mouse was recorded for each day. No significant sex differences were detected. Unpaired two-tailed T-test was performed with Benjamini, Krieger, and Yekutieli's false discovery rate (FDR) method to correct for multiple comparisons.         oligodendrocytes, OPCs, astrocytes, and microglia were purified by immunopanning from whole brains of P17 PD:ibot and littermate control mice. Anti-GalC hybridoma was used to purify differentiated oligodendrocytes; anti-O4 hybridoma was used to isolate OPCs; anti-HepaCAM antibody was used to purify astrocytes; anti-CD45 antibody was used to isolate microglia. More details can be found in the method section. Gene expression was determined by RNA-seq. Genes exhibiting significant  Figure 5E and G). We next quantified the size of lamellar cells, which have large sheaths of myelin-like membrane. Interestingly, we found that lamellar cells from PD:ibot mice are significantly smaller than those from control mice ( Figure 5E and H). We also used a membranestaining version of the CellMask dye at day 7 of differentiation and found similar results (Figure 5figure supplement 1). Together, these observations suggest that VAMP1/2/3-dependent exocytosis is required for the morphological maturation of oligodendrocytes and that exocytosis has a direct effect on cells within the oligodendrocyte lineage to promote their development.

Cell non-autonomous effect of botulinum toxin-B in oligodendrocyte development in vitro
To examine whether cell non-autonomous effect contributes to the oligodendrocyte development defect associated with botulinum toxin-B expression, we compared the growth of wild-type cells in cultures containing vs. not containing botulinum-expressing cells. We took advantage of the fact that all OPCs purified from PD:ibot mice were not botulinum-GFP-expressing (efficiency ~65%, Figure 6A and B). The GFPcells in PD:ibot OPC cultures did not express botulinum toxin and were competent in exocytosis. We compared the development of GFPcontrol cells in cultures generated from PD:ibot mice vs. control cells in cultures generated from control mice. Interestingly, we found that the percentages and the sizes of lamellar cells in control cells from PD:ibot cultures were smaller than in control cells from control cultures ( Figure 6C and D). Although both groups of cells were competent in exocytosis, they were surrounded by exocytosis-deficient vs. exocytosis-competent neighbor cells. The difference in the growth capacity of control cells in the presence of different neighbor cells reveals cell non-autonomous contributions of botulinum-expressing cells in oligodendrocyte development.
Oligodendrocyte-lineage cell-secreted molecules partially restore oligodendrocyte morphological maturation in secretion-deficient cells Cell non-autonomous effects on oligodendrocyte development may be mediated by contactdependent mechanisms or secreted molecules. To distinguish between these possibilities, we next assessed whether adding oligodendrocyte-lineage cell-secreted molecules could restore differentiation in VAMP1/2/3-dependent exocytosis-deficient OPCs. We prepared co-cultures of OPCs separated by inserts with 1 μm-diameter pores to allow for the diffusion of secreted molecules ( Figure 6E). We plated PD:ibot and control cells on inserts and on the bottom of culture wells in four combinations: (1) control-inserts-control-wells; (2) PD:ibot-inserts-control-wells; (3) PD:ibot-inserts-PD:ibotwells; and (4) control-inserts-PD:ibot-wells ( Figure 6F) and examined oligodendrocyte morphological differentiation on the bottom of culture wells by quantifying lamellar cells as described above. We performed one-way ANOVA with a multiple comparison test comparing every group to every other group. Comparing group 3 vs. group 4, we found that adding secreted molecules from control cells on inserts partially rescued the size of lamellar cells of PD:ibot cells on the bottom of culture wells changes (P-value adjusted for multiple comparisons, Padj <0.05, fold change >1.5) are shown in red. (E-H) Examples of mature oligodendrocyte marker gene expression by oligodendrocytes purified from PD:ibot and control mice at P17. N=3 mice per group. Significance is determined by DESeq2. Plp1: 1586±375.9 in control and 729.3±152.4 in PD:ibot, p=3.96 × 10 -13 ; Mbp: 135.2±12.8 in control and 89.8±13.5 in PD:ibot, p=4.76 × 10 -5 ; Aspa: 5.9±1.7 in control and 2.0±0.6 in PD:ibot, p=0.003; Mobp: 274.4±40.1 in control and 97.5±24.0 in PD:ibot, p=0.0004. (I-L) Examples of GO terms associated with genes upregulated in oligodendrocytes (I), downregulated in oligodendrocytes (J), upregulated in OPCs (K), and downregulated in microglia (L). There are no GO terms significantly associated with genes downregulated in OPCs or upregulated in microglia. (M, N) Expression of Ptgds by OPCs and oligodendrocytes purified from PD:ibot and control mice at P17. N=3 mice per group. Significance is determined by DESeq2. Ptgds in OPCs: 11.6±4.3 in control and 34.5±10.5 in PD:ibot, p=1.79 × 10 -12 ; Ptgds in oligodendrocytes: 4.1±0.3 in control and 39.1±13.9 in PD:ibot, p=3.14 × 10 -8 . (O) Expression of Hpgd by microglia purified from PD:ibot and control mice at P17. (E-H, M-O) Expression is shown in RPKM. N=3 mice per group. Significance is determined by DESeq2. Hpgd in microglia: 11.7±0.9 in control and 6.1±0.7 in PD:ibot, p=6.34 × 10 -19 .
The online version of this article includes the following source data and figure supplement(s) for figure 4: Source data 1. It contains original data points for Figure 4E Figure 6G and H). We used both a cytosol-staining version ( Figure 6G) and a membrane-staining version ( Figure 6-figure supplement 1) of the CellMask dye and found similar results. These observations lend further support to the hypothesis that oligodendrocyte-lineage cell-secreted molecules promote oligodendrocyte development.

Blocking L-PGDS leads to oligodendrocyte development defects in vitro
We next sought to uncover the identity of the secreted molecules that promote oligodendrocyte and myelin development. We mined our oligodendrocyte and OPC RNA-seq dataset to identify highly expressed genes encoding secreted proteins (Zhang et al., 2014;Zhang et al., 2016). After testing a few candidate genes using in vitro OPC cultures, we focused on Ptgds, one of the most abundant genes encoding a secreted protein expressed by oligodendrocyte-lineage cells in both humans and mice (Zhang et al., 2014;Zhang et al., 2016). Its expression increases during development (Kang et al., 2011) as oligodendrocyte development and myelination occur. Interestingly, Ptgds is important for Schwann cell myelination in the peripheral nervous system (Trimarco et al., 2014). Yet, its function in the development of the CNS is unknown. Ptgds encodes the L-PGDS protein, which converts prostaglandin H2 to prostaglandin D2 (PGD2) . We performed western blot analyses to assess L-PGDS secretion by botulinum toxin B-expressing OPCs/oligodendrocytes in culture. We detected an increase in intracellular L-PGDS in OPC/oligodendrocyte cultures from PD:ibot mice compared with controls ( Figure 7A). Secreted L-PGDS, however, is lower in PD:ibot compared with control cultures ( Figure 7A-C), suggesting that botulinum toxin B inhibits L-PGDS secretion. L-PGDS secretion is not completely eliminated, most likely because not all cells in the culture express botulinum toxin (efficiency:~65%, Figure 6B) and wild-type cells may compensate by increasing secretion when extracellular L-PGDS levels are low.
To determine the role of L-PGDS in oligodendrocyte development and CNS myelination, we first assessed oligodendrocyte development in vitro in the presence of AT-56, a specific L-PGDS inhibitor (Irikura et al., 2009). We found that AT-56 inhibits wild-type oligodendrocyte development in a dose-dependent manner in vitro ( Figure 7D and E), suggesting a requirement of L-PGDS in oligodendrocyte development, without affecting their survival (Figure 7-figure supplement 1). Using cytoplasm-staining and membrane-staining versions of CellMask dyes, we observed similar results (Figure 7-figure supplement 2).
L-PGDS synthesizes PGD2, whereas another enzyme, 15-hydroxyprostaglandin dehydrogenase (HPGD) inactivates PGD2 (Conner et al., 2001). To assess the involvement of PGD2 in oligodendrocyte development, we added HPGD to OPC cultures from control mice. Interestingly, we found that HPGD reduced the percentage of oligodendrocytes with mature lamellar morphology ( Figure 7F and G), lending further support to the potential role of the L-PGDS/PGD2 pathway in oligodendrocyte development.

L-PGDS is required for oligodendrocyte development in vivo
Having discovered the role of L-PGDS in oligodendrocyte development in vitro, we next assessed whether L-PGDS regulates oligodendrocyte development in vivo. We examined oligodendrocytes in L-PGDS global knockout mice and found a significant decrease in CC1 + oligodendrocytes in the corpus callosum and cerebral cortex of L-PGDS-knockout mice at P9 (Figure 8A and B). The density of OPCs (PDGFRα + ) and cleaved caspase-3 positive cells and the intensity of SMI-32, a marker for damaged axons, and Iba1, a marker for microglial reactivity, did not differ between L-PGDS-knockout and control mice ( Figure 8C-K).
Source data 1. It contains original data points for Figure 6B, C, D and H. Figure 6F.         L-PGDS and PGD2 restore the development of secretion-deficient oligodendrocytes Next, we determined whether L-PGDS is sufficient to rescue the maturation defect of PD:ibot cells in vitro. Indeed, the exogenous addition of recombinant L-PGDS protein partially rescued the percentage of cells with lamellar morphology from PD:ibot mice ( Figure 9A and B), further supporting the role of L-PGDS in oligodendrocyte development. To assess the role of PGD2, the synthesis product of the L-PGDS enzyme, in oligodendrocyte development, we added PGD2 to cultures from PD:ibot mice and observed a partial rescue of the oligodendrocyte maturation defect of PD:ibot cells ( Figure 9C and D). Although other cell types could mediate the effects of systemic L-PGDS knockout, the observation that L-PGDS and PGD2 rescue the morphological maturation of oligodendrocytes in purified cultures in vitro supports a direct role of PGD2 in oligodendrocyte-lineage cell development.

Overexpression of L-PGDS partially rescues the myelin defect in PD:ibot mice in vivo
To assess the effect of L-PGDS on myelin development in vivo, we bred the L-PGDS overexpressing transgenic mouse strain (Ptgds-TG, line B7) (Pinzar et al., 2000) with PD:ibot mice to obtain triple transgenic mice (Pdgfra-CreER; lox-stop-lox-botulinum toxin-B light chain-GFP, Ptgds-TG). Compared with PD:ibot mice, the triple transgenic mice exhibited enhanced myelination ( Figure 9E-H), providing further evidence of the role of L-PGDS in oligodendrocyte and myelin development in vivo. Interestingly, we found that Ptgds is the most highly upregulated gene in OPCs and oligodendrocytes from PD:ibot mice compared with control mice based on RNA-seq ( Figure 4A and B). Moreover, Hpgd, which inactivates PGD2, is the most robustly downregulated gene in microglia from PD:ibot mice compared with control mice ( Figure 4D). These intriguing observations may reflect homeostatic mechanisms that maintain L-PGDS and PGD2 levels. When L-PGDS secretion is blocked by botulinum toxin, extracellular PGD2 levels decrease, which induces an increase in Ptgds mRNA to promote PGD2 production and a decrease in Hpgd mRNA to reduce PGD2 inactivation.
Given the enriched expression of L-PGDS by oligodendrocyte-lineage cells (Zhang et al., 2014;Zhang et al., 2016) and its localization in the extracellular space (Figure 7A and C;Hoffmann et al., 1993), our results indicate that L-PGDS is an oligodendrocyte-lineage cell-secreted autocrine/paracrine molecule that promotes oligodendrocyte development and myelination. Our results are consistent with the following model: OPCs secrete autocrine/paracrine signals such as L-PGDS to promote oligodendrocyte development and myelination. When VAMP1/2/3-dependent exocytosis is blocked, L-PGDS secretion is defective, leading to defective oligodendrocyte development and myelination. Overexpression of Ptgds (L-PGDS) in exocytosis-deficient PD:ibot mice restores L-PGDS levels and partially rescues myelination defect. Our discovery of the role of exocytosis and L-PGDS in oligodendrocytes provides insight into the mechanisms regulating oligodendrocyte development and myelination and reveals novel molecular targets for future efforts aimed toward enhancing myelination for neural repair.

Discussion
In this study, we showed that oligodendrocyte-lineage cell-secreted molecules promote oligodendrocyte development and myelination in an autocrine/paracrine manner. We identified L-PGDS as one such secreted molecule, thus revealing a novel cellular mechanism regulating oligodendrocyte development.
Previously, the roles of VAMP3 and related pathways in myelin protein delivery and oligodendrocyte morphogenesis have been investigated largely in vitro using cultures of an OPC-like cell line (Oli-Neu cells) or primary oligodendrocytes. For example, VAMP3 and VAMP7 knockdown inhibits the transport of a myelin protein, Proteolipid Protein1 in vitro (Feldmann et al., 2011). Tetanus toxin, which cleaves VAMP1/2/3, inhibits oligodendrocyte branching in vitro (Sloane and Vartanian, 2007). Syntaxin4, a potential binding partner of VAMP3, is required for the transcription of MBP in oligodendrocytes in vitro (Bijlard et al., 2015). Our study established a requirement of VAMP1/2/3-dependent exocytosis in oligodendrocyte development, myelination, and motor behavior in vivo and identified L-PGDS as an oligodendrocyte-lineage cell-secreted protein that promotes oligodendrocyte development and myelination. VAMP1/2/3-dependent exocytosis is not the only pathway employed by oligodendrocyte-lineage cells to release molecules that mediate cell-cell interactions. For example, oligodendrocytes release exosome-like vesicles that inhibit the growth of myelin-like membranes in vitro (Bakhti et al., 2011). Tetanus toxin cleaves VAMP1/2/3 but does not affect exosome release (Fader et al., 2009). Therefore, the role of VAMP1/2/3-dependent exocytosis in promoting myelination and the effect of exosomelike vesicles in inhibiting myelination are likely parallel pathways independent of each other. In future studies, it could be interesting to determine the signals that regulate VAMP1/2/3-dependent exocytosis and VAMP1/2/3-independent exosome release during development and disease in vivo and thus define how these two seemingly opposing effects are coordinated to shape precise and dynamic myelination.
We observed a decrease in the percentage of PDGFRa + OPCs expressing botulinum toxin-B from P8 to P30 in PD:ibot mice ( Figure 1G and Figure 1-figure supplement 2C). Since OPC proliferation is increased and the overall OPC densities do not change in PD:ibot mice, a decrease in the percentage of OPCs expressing the toxin is consistent with the death of some toxin-expressing OPCs. Although our cleaved caspase-3 immunohistochemistry results did not show a difference in OPC survival between PD:ibot and control cells (Figure 3-figure supplement 1), we cannot exclude the possibilities that OPC die through non-apoptotic mechanisms or that microglia clear dead cells too rapidly for accurate counting. The online version of this article includes the following source data for figure 9: Source data 1. It contains original data points for Figure 9B, D, F, G and H.

Figure 9 continued
OPCs are present throughout the CNS in adults (Hughes et al., 2013;Kang et al., 2010), even in demyelinated lesions in patients with multiple sclerosis (Franklin, 2002). Therefore, inducing oligodendrocyte and new myelin formation is an attractive strategy for treating demyelinating diseases. However, remyelination therapy has not been successful so far (Franklin, 2002), underscoring the need for a more complete understanding of the mechanisms regulating oligodendrocyte and myelin development. Our discovery of the role of L-PGDS in oligodendrocyte development and myelination adds to the knowledge of the molecular regulation of myelination. Interestingly, mixed results have been reported on the level of the Ptgds gene and the L-PGDS protein in multiple sclerosis patients and mouse models (Jäkel et al., 2019;Kagitani-Shimono et al., 2006;Penkert et al., 2021). During remyelination in mice, PGD2 levels increase (Penkert et al., 2021). Future studies should clarify the involvement of L-PGDS in multiple sclerosis and its therapeutic potential in promoting remyelination.
A recent study shows that the gene encoding L-PGDS, Ptgds, marks a subpopulation of OPCs more resilient to spinal cord injury than other OPCs (Floriddia et al., 2020). Thus, the function of L-PGDS in OPC heterogeneity and their responses to injury and other neurological disorders will be interesting to explore in the future.
The product of the L-PGDS enzyme, PGD2, binds and activates two G-protein-coupled receptors, DP1 and DP2 (Gpr44) (Narumiya and Furuyashiki, 2011). In addition, PGD2 undergoes nonenzymatic conversion to 15d-PGJ2, which activates the peroxisome proliferator-activated receptor-γ (Scher and Pillinger, 2005). Future studies should aim to identify the receptor(s) that mediates the effect of L-PGDS on oligodendrocyte development and myelination, as well as the downstream signaling pathways.
OPCs and oligodendrocytes express numerous genes encoding secreted proteins (Zhang et al., 2014). Although we identified the role of L-PGDS in oligodendrocyte development, our results do not rule out contributions from other secreted molecules. Our RNA-seq dataset provides a roadmap for future investigation of the roles of additional oligodendrocyte-lineage cell-secreted molecules in the brain.
Blocking exocytosis with botulinum toxin B may reduce the delivery of proteins and lipids to the plasma membrane, therefore causing cell-autonomous effects on oligodendrocyte development in addition to blocking secretion. Both cell-autonomous and cell-non-autonomous mechanisms may be involved in the effect of blocking exocytosis on oligodendrocyte development (Fekete et al., 2022;Lam et al., 2022). Our transwell rescue experiment and the comparisons between control cells in PD:ibot culture and control cells in control culture ( Figure 6) support the importance of secreted molecules but do not rule out cell-autonomous mechanisms.
Two recent publications from the Zuchero and Nishiyama groups also utilized the ibot mice to investigate the role of VAMP2/3-mediated exocytosis in oligodendrocyte development and myelination (Fekete et al., 2022;Lam et al., 2022). Together with data presented here, the three complementary studies utilized different Cre lines to express botulinum toxin-B in oligodendrocyte lineage cells (NG2-Cre in Fekete et al., CNP-Cre in Lam et al., and PDGFRa-CreER in this study). Interestingly, all three ibot mouse strains exhibit myelination defects in the central nervous system (the spinal cord in Fekete et al., the brain in this study, and the brain and the spinal cord in Lam et al.). These consistent results from three mouse strains obtained by independent groups demonstrate an important role of VAMP2/3-mediated exocytosis in myelin development. VAMP2/3-mediated exocytosis may affect membrane expansion, secretion, and signaling. To provide mechanistic insight into the function of VAMP2/3 in myelin development, these three studies focused on different aspects of VAMP2/3 function. In NG2:ibot mice, Fekete et al. observed elevated levels of Fyn kinase, which is a signaling molecule implicated in regulating oligodendrocyte maturation and myelination. Using CNP:ibot mice, Lam et al. demonstrated the role of VAMP2/3-dependent exocytosis in membrane expansion and insertion of plasma membrane proteins during oligodendrocyte maturation, and further performed elegant imaging studies to capture exocytosis events of VAMP2/3-containing vesicles in oligodendrocytes in vitro and in vivo. In this study, we demonstrated the cell-non-autonomous function of VAMP2/3-dependent exocytosis in oligodendrocyte development and identified a novel role of a secreted molecule, L-PGDS, in oligodendrocyte development. Because VAMP2/3-mediated exocytosis may contribute to the membrane insertion and secretion of a variety of molecules, which may in turn activate multiple intracellular signaling pathways, it is not surprising that multiple mechanisms mediate the effect of VAMP2/3 in oligodendrocyte development and myelination. These three complementary studies provide key insight toward a comprehensive understanding of the role of VAMP2/3 in myelination.
Although all three ibot mouse strains exhibit myelin defects, other phenotypes differ among the strains. For example, the ibot mice in Fekete et al. and this study have reduced oligodendrocyte densities, whereas Lam et al. did not observe an overt loss of oligodendrocytes. We used the CC1 antibody to detect differentiated oligodendrocytes and Fekete et al. used CC1 (i.e. QKI7) and several additional markers to validate a reduction in mature oligodendrocytes (ASPA, Nkx2.2). Differences in the timing of botulinum toxin-B expression driven by the Cre lines may account for the divergent results. both Cre lines drive toxin expression in OPCs, whereas Lam et al. used CNP-Cre that predominantly drives toxin expression in oligodendrocytes. The intriguing divergent phenotypes suggest the possibility of a stage-dependent requirement of VAMP2/3-mediated exocytosis in the development of oligodendrocyte-lineage cells, which may be investigated in the future to clarify the complex roles of VAMP2/3 in oligodendrocyte development and myelination. Together, the three complementary studies revealed the importance of VAMP2/3-mediated exocytosis in myelination through a combination of cell-autonomous and cellnon-autonomous mechanisms.

Lead contact and materials availability
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Ye Zhang (yezhang@ucla.edu). This study did not generate new unique reagents.

Experimental animals
All animal experimental procedures (protocols: #R-16-079 and #R-16-080) were approved by the Chancellor's Animal Research Committee at the University of California, Los Angeles, and conducted in compliance with national and state laws and policies. All the mice were group-housed in standard cages (maximum 5 mice per cage). Rooms were maintained on a 12 hr light/dark cycle. Pdgfra-CreER (Jax #018280), ibot (Jax #018056), Gfap-Cre (Jax #024098), and Tek-Cre (Jax, #004128) mouse strains were obtained from Jackson Laboratories. L-PGDS global knockout (Ptgds -/-) mouse strain was originally from Urade, cryopreserved by Garret FitzGerald . L-PGDS overexpressing transgenic mouse strain (Ptgds-TG, line B7) was originally from Urade, cryopreserved by JCRB Laboratory Animal Resource Bank (Pinzar et al., 2000). All PD:ibot and control mice, including wildtype, Cre-only and ibot-only mice used were congenic. We used PD:ibot and littermate control mice raised under the same maternal care in the same cage to minimize environmental confounding factors.

OPC purification and culture
Whole brains excluding the olfactory bulbs and the cerebellum from one pup at postnatal day 7 to day 8 were used to make each batch of OPC culture. OPCs were purified using an immunopanning method described before (Emery and Dugas, 2013). Briefly, the brains were digested into single-cell suspensions using papain. Microglia and differentiated oligodendrocytes were depleted using anti-CD45 antibody-(BD Pharmingen, cat #550539) and GalC hybridoma-coated panning plates, respectively. OPCs were then collected using an O4 hybridoma-coated panning plate. For most culture experiments, cells were plated on 24-well plates at a density of 30,000 per well. For comparison of OPC differentiation at different densities, OPCs were plated at densities of 5000 per well and 40,000 per well. For all experiments, OPCs were first kept in proliferation medium containing growth factors PDGF (10 ng/ml, Peprotech, cat #100-13 A), CNTF (10 ng/ml, Peprotech, cat #450-13), and NT-3 (1 ng/ml, Peprotech, cat #450-03) for 2-3 days, and then switched to differentiation medium containing thyroid hormone (40 ng/ml, Sigma, cat #T6397-100MG) but without PDGF or NT-3 for seven days to differentiate them into oligodendrocytes as previously described (Emery and Dugas, 2013). Half of the culture media was replaced with fresh media every other day. All the cells were maintained in a humidified 37°C incubator with 10% CO 2 . Cells from both female and male mice were used. For coculture experiments with inserts, OPCs were purified from PD:ibot and littermate control mice as described above. A total of 100,000 cells per well were plated on inserts with 1 μm diameter pores (VWR, cat #62406-173), and the inserts were placed on top of wells with cells plated at 30,000 cells per well density on 24-well culture plates. 200 μl medium was added per insert and 500 μl medium was added per well under the inserts.

Drugs and treatment
4-Hydroxy-tamoxifen stock solutions were made by dissolving 4-hydroxy-tamoxifen (Sigma, H7904) into pure ethanol at 10 mg/ml. The stock solutions were stored at -80°C until use. On the day of injection, an aliquot of 4-hydroxy-tamoxifen stock solution (100 μl) was thawed and mixed with 500 μl sunflower oil by vortexing for 5 min. Ethanol in the solution was vacuum evaporated in a desiccator (VWR, for an hour. 0.1 mg 4-hydroxy-tamoxifen was injected into each mouse subcutaneously daily for 2 days at P2 and P4. An L-PGDS inhibitor, AT-56 (Cayman Chemicals, cat #13160), and prostaglandin D2 (Cayman Chemicals, cat #12010) were dissolved in dimethyl sulfoxide (DMSO). To inhibit L-PGDS activity in vitro, AT-56 was added to the oligodendrocyte culture medium at 1 µM and 5 µM every other day. HPGD protein (R&D system, cat#5660-DH-010) was added to the oligodendrocyte culture medium at 3 µM and 6 µM daily. L-PGDS protein (Cayman Chemicals, cat#10006788) was added to the oligodendrocyte culture at 20 µM daily. For prostaglandin D2 treatment, prostaglandin D2 was added to the oligodendrocyte culture medium at 1 µM and 2 µM every 12 hr. An equal amount of DMSO was added to the control wells. Because a metabolite of prostaglandin D2, 15-d-prostaglandin J2, induces cell death, which can be prevented by N-acetyl cysteine (Lee et al., 2008), we included 1 mM N-acetyl cysteine, which is shown to improve cell survival, in the culture media of prostaglandin D2-treated and control cells.

Purification of microglia, oligodendrocytes, oligodendrocyte precursor cells, and astrocytes
Whole brains excluding the olfactory bulbs and the cerebellum from P17 PD:ibot and control littermates were used for the purification of microglia, oligodendrocytes, oligodendrocyte precursor cells, and astrocytes. A single-cell suspension was prepared as described above. Cells were incubated for 30 min on an anti-CD45 antibody (BD Pharmingen, cat#550539, 1.25 µg/ml)-coated panning plate to harvest microglia, followed by two sequential CD45-coated panning plates to deplete remaining microglia. Cell suspension was then incubated for 30 min on a GalC hybridoma-coated panning plate to collect differentiated oligodendrocytes, followed by two more GalC hybridoma-coated plates to deplete any remaining differentiated oligodendrocytes. The remaining cells were incubated for 30 min on an O4 hybridoma-coated panning plate to collect oligodendrocyte precursor cells, followed by two O4 hybridoma-coated panning plates to deplete remaining OPCs. The cell suspension was then incubated with an anti-HepaCAM antibody (R&D Systems, cat# MAB4108)-coated panning plate to collect astrocytes.

RNA-seq
Total RNA was extracted using the miRNeasy Mini kit (Qiagen cat #217004). The concentrations and integrities of the RNA were measured using TapeStation (Agilent) and Qubit. Sixty ng total RNA from each sample was used for library preparation. cDNA was generated using the Nugen Ovation V2 kit (Nugen) and fragmented using the Covaris sonicator. Sequencing libraries were prepared using the Next Ultra RNA Library Prep kit (New England Biolabs) with 12 cycles of PCR amplification. An Illumina HiSeq 4000 sequencer was used to sequence these libraries and each sample had an average of 19.1 ± 2.9 million 50 bp single-end reads.

RNA-seq data analysis
STAR package was used to map reads to mouse genome mm10 and HTSEQ was used to obtain raw counts from sequencing reads. EdgeR-Limma-Voom packages in R were used to calculate Reads per Kilobase per Million Mapped Reads (RPKM) values from raw counts. DESeq2 package was used to analyze differential gene expression.
A large area of cerebral cortex, including motor cortex and somatosensory cortex, from bregma 1 mm to -2 mm was used for the following quantification: MBP coverage in the cortex, the density of oligodendrocyte-lineage marker genes (CC1, Olig2, PDGFRα, Enpp6, and Mbp), and the density of apoptotic cells (cleaved caspase-3 + ).
For immunocytochemistry of cultured cells, cells were fixed with 4% PFA and 0.3% Tween-20 in PBS. After blocking in 5% donkey serum, cells were then stained with the primary antibodies described above, VAMP2 (Synaptic Systems, cat #104 211, dilution 1:100), and VAMP3 (Synaptic Systems, cat #104 103, dilution 1:500) at 4 °C overnight. After three washes in PBS, cells were stained with secondary antibodies and the cytoplasm-staining CellMask Blue (Invitrogen, cat #H32720, 1:1,000) at 4 °C overnight. To stain the cells with the membrane-staining CellMask (Invitrogen, cat#C37608 Dilution 1:1000), we added the dye into the culture medium and incubate it with the cells for 13 min at 37 °C. After staining, cells were washed once with PBS and fixed with 4% PFA for 15 min at room temperature. Cells were washed three times with PBS before covered with mounting medium (Fisher, cat #H1400NB). Slides were imaged with a Zeiss Apotome epifluorescence microscope.
Fluorescence microscopy images were cropped, and brightness contrast was adjusted with identical settings across genotype, treatment, and control groups using Photoshop and ImageJ. All the images were randomly renamed using the following website (https://www.random.org/) and quantified with the experimenter blinded to the genotype and treatment condition of the samples. Cells with MBP + membrane spreading out were identified as lamellar cells. Illustrations were made with Biorender.

RNAscope in situ hybridization
P15 mice were transcardially perfused with PBS followed by 4% PFA. Brains were removed and postfixed in 4% PFA for an additional 2 hr at room temperature and then overnight at 4 °C. Tissues were then dehydrated in 30% sucrose, embedded in OCT compound, and cut into 20-to 30-μm-thick sections. RNAscope Multiplex Fluorescent Reagent Kit v2 (ACDBio, cat# 323100) was used per manufacturer's protocol. Probes used in this paper were purchased from ACDBio for mouse Enpp6 (cat#511021-C2) and Mbp (cat# 451491). Slides were imaged with a Zeiss Apotome epifluorescence microscope at equal power and exposure across all samples stained with the same set of probes.

Transmission electron microscopy
Brain specimens for transmission electron microscopy were prepared as described before (Salazar et al., 2018). Mice were anesthetized using isoflurane and transcardially perfused with 0.1 M phosphate buffer (PB) followed by 4% PFA with 2.5% glutaraldehyde in 0.1 M PB buffer. Brains were removed and post-fixed in 4% PFA with 2.5% glutaraldehyde in 0.1 M PB for another two days. Brains were sliced with Young Mouse Brain Slicer Matrix (Zivic Instruments, cat #BSMYS001-1) and a small piece of the corpus callosum was isolated from brain sections at 0-1 mm anterior to Bregma. After washing, samples were then post-fixed in 1% osmium tetroxide in 0.1 M PB (pH 7.4) and dehydrated through a graded series of ethanol concentrations. After infiltration with Eponate 12 resin, samples were embedded in fresh Eponate 12 resin and polymerized at 60 °C for 48 hr. Ultrathin sections of 70 nm thickness were prepared and placed on formvar/carbon-coated copper grids and stained with uranyl acetate and lead citrate. Grids were examined using a JEOL 100 CX transmission electron microscope at 60 kV and images were captured by an AMT digital camera (Advanced Microscopy Techniques Corporation, model XR611) by the Electron Microscopy Core Facility, UCLA Brain Research Institute. Before analysis, TEM images were blinded as described above. An axon that is encircled by 2 or more compacted layers of electron-dense lines is defined as a myelinated axon. g-ratio was determined by dividing the mean diameter of the area inside myelin by the mean diameter of the same axon with myelin. Approximately 100 myelinated axons from each group were analyzed for g-ratio analysis. For measuring the percentage of myelinated axons, approximately 900 axons from three mice were used for each group. More than 1300 axons from four mice from each group were used to measure axon density and approximately 500 axons from four mice from each group were used for axon diameter measurement.

RNA extraction and qPCR
100 k Day 7 differentiated oligodendrocytes were lysed with the Qiazol lysis buffer and RNA was extracted using the PureLink RNA mini kit (Invitrogen 12183018 A) following the manufacturer's protocol. cDNA was generated using the SuperScript III First-Strand Synthesis SuperMix (Invitrogen 18080400). The primer for Mbp was designed using the primerblast tool from NCBI. Primerblast was used to validate the specificity of primers. The sequences for Mbp and Gapdh primers were provided in the key resource table. PowerUp SYBR Green Master Mix (Applied Biosystems A25742) and a QuantStudio 3 Real-Time PCR System (Thermo fisher, cat# A28567) were used for qRT-PCR reaction.

Western blot
We purified OPCs from PD:ibot and control mice by immunopanning and cultured them in proliferation medium for 2-3 days and differentiation medium for 7 days as described above. To collect secreted samples, culture media were mixed with ethylenediaminetetraacetic acid (EDTA)-free protease inhibitor cocktail (Sigma, cat #4693159001) at a 6:1 ratio and centrifuged at 1000×g for 10 min to remove dead cells and debris. To collect whole-cell lysates, cells were washed with PBS, lysed with radioimmunoprecipitation assay buffer containing EDTA-free protease inhibitor cocktail, and centrifuged at 12,000×g for 10 min to remove cell debris.

Motor behavior
Mice were familiarized with being picked up and handled by the experimenter daily for three days before the test to reduce stress. Mice were also habituated to the rotarod testing room for 15 min prior to all testing. Both male and female adult mice (2-5 months old) were used in the rotarod test. Mice were given three trials per day for three consecutive days (5-60 rpm over 5 min, with approximately 30 min between successive trials). The latency to fall was measured and the experimenter was blinded to the genotype of the mice during the test.

Quantification and statistical analysis
The numbers of animals and replicates are described in the figures and figure legends. The RNA-seq data were analyzed using the DESeq2 package. Adjusted p-values smaller than 0.05 were considered significant. For all non-RNA-seq data, analyses were conducted using Prism 8 software (Graphpad).
The normality of data was tested by the Shapiro-Wilke test. For data with a normal distribution, Welch's t-test was used for two-group comparisons and one-way ANOVA was used for multi-group comparisons. An estimate of variation in each group is indicated by the standard error of the mean (S.E.M.). * p<0.05, ** p<0.01, *** <0.001. An appropriate sample size was determined when the study was being designed based on published studies with similar approaches and focus as our study. A biological replicate is defined as one mouse. Different culture wells from the same mouse or different images taken from the brains or cell cultures from the same mouse are defined as technical replicates. All statistical tests were performed with each biological replicate/mouse as an independent observation. The number of times each experiment was performed is indicated in figure legends. No data were excluded from the analyses. Mice and cell cultures were randomly assigned to treatment and control groups. Imaging analyses and behavior tests were conducted when the experimenter was blinded to the genotypes or treatment conditions.

Code availability
This study did not generate new codes.
OPCs, oligodendrocytes, microglia, and astrocytes from PD:ibot and littermate control mice at P17. No gene ontology terms were significantly enriched in genes downregulated in OPCs, astrocytes, or upregulated in microglia in PD:ibot mice.

Data availability
We deposited all RNA-seq data to the Gene Expression Omnibus under accession number GSE168569.
The following dataset was generated: