key: cord-0288162-offspfhd
authors: Merour, Emeric; Hmidan, Hatem; Marie, Corentine; Helou, Pierre-Henri; Lu, Haiyang; Potel, Antoine; Hure, Jean-Baptiste; Clavairoly, Adrien; Shih, Yi Ping; Goudarzi, Salman; Dussaud, Sebastien; Ravassard, Philippe; Hafizi, Sassan; Lo, Su Hao; Hassan, Bassem A.; Parras, Carlos
title: Transient regulation of focal adhesion via Tensin3 is required for nascent oligodendrocyte differentiation
date: 2022-02-28
journal: bioRxiv
DOI: 10.1101/2022.02.25.481980
sha: 2d8e2e97680ef6941c1296a2d35b0512eae7ab23
doc_id: 288162
cord_uid: offspfhd

The differentiation of oligodendroglia from oligodendrocyte precursor cells (OPCs) to complex and extensive myelinating oligodendrocytes (OLs) is a multistep process that involves largescale morphological changes with significant strain on the cytoskeleton. While key chromatin and transcriptional regulators of differentiation have been identified, their target genes responsible for the morphological changes occurring during OL myelination are still largely unknown. Here, we show that the regulator of focal adhesion, Tensin3 (Tns3), is a direct target gene of Olig2, Chd7, and Chd8, transcriptional regulators of OL differentiation. Tns3 is transiently upregulated and localized to cell processes of immature OLs, together with integrin-β1, a key mediator of survival at this transient stage. Constitutive Tns3 loss-of-function leads to reduced viability in mouse and humans, with surviving knockout mice still expressing Tns3 in oligodendroglia. Acute deletion of Tns3 in vivo, either in postnatal neural stem cells (NSCs) or in OPCs, leads to a two-fold reduction in OL numbers. We find that the transient upregulation of Tns3 is required to protect differentiating OPCs and immature OLs from cell death by preventing the upregulation of p53, a key regulator of apoptosis. Altogether, our findings reveal a specific time window during which transcriptional upregulation of Tns3 in immature OLs is required for OL differentiation likely by mediating integrin-β1 survival signaling to the actin cytoskeleton as OL undergo the large morphological changes required for their terminal differentiation.

Oligodendrocyte lineage cells, mainly constituted by oligodendrocyte precursor cells (OPCs) and oligodendrocytes (OLs), play key roles during brain development and neuronal support, by allowing saltatory conduction of myelinated axons and metabolically supporting these axons with lactate or glucose shuttling through the myelin sheath (Funfschilling et al., 2012; Lee et al., 2012; Meyer et al., 2018) . Accumulating evidence also indicates their fundamental contribution to different aspects of adaptive myelination, a type of brain plasticity (Mount and Monje, 2017; Yang et al., 2020) , shown by the requirement of new oligodendrogenesis for proper learning and memory in motor, spatial, and fear-conditioning learning paradigms (McKenzie et al., 2014; Xiao et al., 2016; Steadman et al., 2019; Pan et al., 2020; Wang et al., 2020; Xin and Chan, 2020) . Furthermore, oligodendroglia and myelin pathologies have been recently linked, not only to the development of glioma (Liu et al., 2011) , but to developmental (Castelijns et al., 2020; Phan et al., 2020) , neurodegenerative (Grubman et al., 2019; Bryois et al., 2020) , and psychiatric (Nott et al., 2019) diseases.

Unlike most precursor cells, OPCs constitute a stable population of the postnatal and adult CNS (Ffrench-Constant and Raff, 1986; Suzuki and Goldman, 2003) . Therefore, OPCs need to keep a tight balance between proliferation, survival, and differentiation. This balance is crucial to maintain the OPC pool while contributing to myelin plasticity in adult life, and to remyelination in diseases such as multiple sclerosis (MS). Demyelinated MS plaques can be normally repaired in early stages of the disease, presumably by endogenous OPCs, but this repair process becomes increasingly inefficient with aging, when OPC differentiation seems to be partially impaired (Chang et al., 2002; Compston and Coles, 2002; Neumann et al., 2019) . Therefore, understanding the mechanisms involved in OPC differentiation is critical to foster successful (re)myelination in myelin pathologies.

A large diversity of extrinsic signals, including those mediated by integrin signaling (reviewed in (Bergles and Richardson, 2016) ) as well as many intrinsic factors, including transcription factors (TFs) and chromatin remodelers (reviewed in (Emery and Lu, 2015; Parras et al., 2020) ), are involved in OPC proliferation, survival, and differentiation. However, the mechanisms for how these signals balance OPC behavior is largely unknown. OPC differentiation requires profound changes in chromatin and gene expression (Emery and Lu, 2015; Küspert and Wegner, 2016; Wheeler and Fuss, 2016) . TFs, such as Olig2, Sox10, Nkx2.2 or Ascl1, are key regulators of OL differentiation by directly controlling the transcription of genes implicated in this process (Qi et al., 2001; Stolt et al., 2002; Nakatani et al., 2013; Yu et al., 2013) but being already expressed at the OPC stage, it is still unclear how these TFs control the induction of differentiation. A growing body of evidence suggests that some of these TFs work together with chromatin remodeling factors during transcriptional initiation/elongation to drive robust transcription (Zaret and Mango, 2016) . Accordingly, Olig2 and Sox10 TFs have been shown to cooperate with chromatin remodelers such as Brg1 (Yu et al., 2013) , Chd7 (He et al., 2016; Marie et al., 2018) , Chd8 Zhao et al., 2018) , and EP400 (Elsesser et al., 2019) , to promote the expression of OL differentiation genes. To improve our understanding of the mechanisms of OL differentiation, we searched for novel common targets of these key regulators, by generating and analyzing the common binding profiles of Olig2, Chd7, and Chd8, in gene regulatory elements of differentiating oligodendroglia. We identified Tns3, coding for the focal adhesion protein Tensin3, as one such target and showed that it is expressed in immature OLs during myelination and remyelination, thus constituting a marker for this transient oligodendroglial stage. Using different genetic strategies to induce Tns3 loss-of-function mutations in vivo, we describe, the function of a Tensin family member in the CNS, demonstrating that Tns3 is required for oligodendrocyte differentiation in the postnatal mouse brain, at least in part by mediating integrin-β1 signaling, essential for survival of differentiating oligodendroglia (Colognato et al., 2002; Benninger et al., 2006) .

To find new factors involved in oligodendrocyte differentiation, we screened for target genes of Olig2, Chd7, and Chd8, key regulators of oligodendrogenesis (Lu et al., 2000; Lu et al., 2002; Yu et al., 2013; He et al., 2016; Marie et al., 2018; Zhao et al., 2018; Parras et al., 2020) . We generated and compared the genome-wide binding profiles for these factors in acutely purified oligodendroglial cells from postnatal mouse brain cortices by magnetic cell sorting (MACS) of O4 + cells . MACS-purified cells, composed of 80% PDGFRα + OPCs and 20% of Nkx2.2 + /CC1 + immature oligodendrocytes (iOLs), were subjected to chromatin immunoprecipitation followed by sequencing (ChIP-seq) for Olig2 and histone modifications marking the transcription activity of gene regulatory elements (H3K4me3, H3K4me1, H3K27me3, and H3K27ac; Fig. 1A ). The profile of activity histone marks at Olig2 binding sites indicated that Olig2 binds promoters (60%) and enhancers (40%) with either active or more poised/repressive states ( Fig. S1A-F) , supporting the suggested pioneer function of Olig2 in oligodendrogenesis (Yu et al., 2013) . Among the 16,578 chromatin sites bound by Olig2 corresponding to 8,672 genes (Fig. S1D) , there were key regulators of oligodendrocyte differentiation, including Ascl1, Sox10, Myrf, Chd8, and Smarca4/Brg1 (Fig. 1B; supplementary table 1). Combining Olig2 with Chd7 and Chd8 binding profiles, that we previously generated using the same protocol , we found 1774 chromatin sites commonly bound by the three regulators, with half of them (47% and 832 peaks) corresponding to active promoters (H3K4me3/H3K27ac marks) of 654 protein-coding genes (Fig. 1C, supplementary table 2). Among these genes, Tns3 coding for Tensin3, a focal adhesion protein deregulated in certain cancers (Martuszewska et al., 2009) , showed the highest expression levels in iOLs relative to other brain cell types (Zhang et al., 2014; Fig. 1D) . Indeed, Olig2, Chd7, and Chd8 commonly bound three putative promoters of Tns3 having active transcription marks in purified oligodendroglia (H3K27ac/H3K24me3; Fig. 1E ). To directly assess whether Tns3 expression requires the activity of these key regulators, we interrogated the transcriptomes of these oligodendroglial cells purified from Chd7iKO (Pdgfra-CreER T ; Chd7 flox/flox ), Chd8cKO (Olig1 Cre ; Chd8 flox/flox ), and their respective control cortices Zhao et al., 2018) . Tns3 transcripts were largely downregulated upon acute deletion of these factors in postnatal OPCs/iOLs (Fig. 1F, G) , indicating that Tns3 expression in OPCs/iOLs is directly controlled by Chd7 and Chd8 chromatin remodelers, key regulators of oligodendrocyte differentiation.

We then investigated Tns3 expression pattern in the brain. High expression levels of Tns3 transcripts in iOLs, compared to its low expression in other cells of the postnatal mouse brain detected by bulk transcriptomics (Fig S2A; Zhang et al., 2014) were paralleled by the sparse labelling with Tns3 probes enriched in the white matter of the postnatal and adult brain detected by in situ hybridization ( Fig. S2B ; Allen Brain atlas, https://portal.brain-map.org/. By harnessing single-cell transcriptomics (scRNA-seq), we sought to create an integrative gene profiling for oligodendrocyte lineage cells, by bioinformatics integration and analyses of oligodendrocyte lineage cells at embryonic, postnatal, and adult stages (Marques et al., 2016 (Marques et al., , 2018 . We thus integrated these datasets using Seurat (Stuart et al., 2019) and selected 5,516 progenitor and oligodendroglial cells. Unsupervised clustering and visualization of cells in two dimensions with uniform manifold approximation and projection (UMAP), identified nine different clusters following a differentiation trajectory. Based on known cell-subtype specific markers ( Fig. S2E and supplementary Table3), we could identify these clusters as (Fig. S2C ,E):

(1) two types of neural stem/progenitor cells, that we named NSCs and NPCs according to their expression of stem cell (Vim, Hes1, Id1) and neural progenitor (Sox11, Sox4, Dcx) markers; (2) OPCs expressing their known markers (Pdgfra, Cspg4, Ascl1) and cycling OPCs also enriched in cell cycle markers (Mki67, Pcna, Top2); (3) two stages of iOLs, both expressing the recently proposed markers Itpr2 and Enpp6 (Marques et al., 2016; Xiao et al., 2016) , and which are split by the expression of Nkx2-2 (iOL1 being Nkx2-2 + and iOL2 being Nkx2-2 -), in agreement with our previous characterization by immunofluorescence (Nakatani et al., 2013; Marie et al., 2018) ; (4) myelin forming oligodendrocytes (MFOLs), enriched in markers such as Slc9a3r2 and Igsf8; and (5) two clusters of myelinating OLs, that we named MOL1 and MOL2, expressing transcripts of myelin proteins (Cnp, Mag, Mbp, Plp1, Mog) and some specific markers of each cluster, including Mgst3, Pmp22 for MOL1 (corresponding to MOL1/2/3/4 clusters of Marques 2016), and Neat1, Grm3, Il33 for MOL2 (corresponding to MOL5/6 clusters of Marques 2016). Interestingly, Tns3 transcripts were strongly expressed in both iOL1 and iOL2 clusters (Fig. S2D ,E), similar to the recently proposed iOL markers Itpr2 and Enpp6 (Fig.   S2E ), and downregulated in mature/terminally differentiated oligodendrocytes, indicating that high levels of Tns3 expression are specific to iOLs. We finally, assessed whether Tns3 expression pattern was conserved in human oligodendroglia pursuing a similar bioinformatics analysis using single cell transcriptomes from human oligodendroglia differentiated from embryonic stem cells (Chamling et al., 2021) . Upon integration with Seurat and identification of cluster cell types using specific markers, we selected 7,690 progenitor and oligodendroglial cells that corresponded to six main cluster cell types following a differentiation trajectory from neural cells (NSCs) up to immature OLs (iOL1 and iOL2), as depicted by UMAP representation (Fig. S2F,H) . Cells expressing high levels of TNS3 corresponded to immature oligodendrocytes (iOL1 and iOL2 clusters, Fig. S2G,H) , indicating a conserved expression pattern of Tns3/TNS3 between mouse and human oligodendrogenesis.

Given the high expression level of Tns3 transcripts in immature oligodendrocytes, we characterized the Tns3 protein expression pattern in the postnatal brain using commercial and homemade Tns3-recognizing antibodies. Optimization of immunofluorescence protocols demonstrated signal in CC1 + oligodendrocytes in the postnatal brain with four different antibodies (P24, Fig. S3 ). To our surprise, while all antibodies showed signal localized in the cytoplasm and main processes of CC1+ OLs (Fig. S3A-E) , one Tns3-recognizing antibody (Millipore) also presented a strong nuclear signal never reported for Tns3 localization in other tissues (such as lung, liver, and intestine) (Katz et al., 2007; Nishino et al., 2012; Cao et al., 2015) . To better characterize Tns3 protein expression pattern and its subcellular localization, we generated a knock-in mice tagging the Tns3 C-terminal side with a V5-tag (Tns3 Tns3-V5 mice) by microinjecting mouse zygotes with a single strand oligodeoxynucleotide (ssODN) containing V5 sequence together with Cas9 protein and a gRNA targeting the stop codon region of Tns3 (Methods; Fig. S4A -C). We first verified by immunofluorescence that Tns3-V5 protein in Tns3 Tns3-V5 mice presented the expression pattern reported for Tns3 in the lung and the kidney (Fig. S4D ,E). We then characterized Tns3 protein expression in oligodendroglia using V5 antibodies, finding that Tns3 protein can be detected at high levels in the cytoplasm and main processes of CC1 + iOLs but not in their nuclei ( Fig. 2A) . Using an antibody recognizing Itpr2, a suggested iOL marker (Marques et al., 2016) , we saw that Tns3 largely overlapped with Itpr2 ( Fig. 2B ). Using Nkx2.2 and Olig1 cytoplamic expression distinguishing iOL1 and iOL2 respectively, we found high levels of Tns3 in iOL1s (Nkx2.2 + /Olig1cells) and a fraction of iOL2s (Nkx2.2 -/Olig1 cytoplamic cells; Fig. 2C ), suggesting that Tns3 protein expression peaks in early iOLs. Comparison with Opalin protein localized in the cell body, processes, and myelin segments of oligodendrocytes, showed that Tns3 levels decreased with increasing levels of Opalin, with Tns3-V5 levels undetectable in myelinating oligodendrocytes (i.e. Opalin + /CC1 + cells presenting myelinated segments; Fig. 2D , arrowheads). We then performed Western blot analysis with anti-V5 antibodies in purified O4 + cells from P7, P14, and P21 Tns3 Tns3-V5 mouse brains to assess their specificity to recognize Tns3-V5, knowing that two Tns3 isoforms can be detected at the transcript level in the human brain (Fig. S4F , GTEX project, gtexportal.org/home/gene/TNS3). Indeed, we could detect both the full-length (1450 aa, 155 kDa) and the Tns3 short (C-term, 61 kDa) isoforms in O4+ cells from brains at P7 and P14 stages having many iOLs, but not at P21 having mainly mOLs (Fig. S4G,H) , thus validating the specificity of the anti-V5 antibodies in recognizing Tns3 protein. We eventually found a Tns3 antibody also recognizing the C-terminal of Tns3 protein (Sigma Ct) that upon optimized immunofluorescence labeling confirmed the Tns3 expression pattern seen with the V5 antibodies. In combination with Nkx2.2 and Olig1 immunofluorescence, it showed that Tns3 is strongly detected in the cytoplasm and main cellular processes of all iOL1s, defined as Nkx2.2 high /Olig1cells having a round nucleus and small cytoplasm (Fig. 2E , white arrows), and it divided iOL2s, defined as Nkx2.2 -/CC1 high cells, into three stages: (1) Tns3 high /Nkx2.2 -/Olig1 - (Fig. 2E , arrowheads), (2) Tns3 high /Nkx2.2 -/Olig1 high-cytoplamic (Fig. 2E , grey arrows), and (3) Tns3 -/Nkx2.2 -/ Olig1 high-cytoplamic (Fig. 2E) . A similar Tns3 expression pattern and localization was found in vitro using neonatal neural progenitors' differentiating cultures, where Tns3 was detected together with CNP myelin protein in the cytoplasm and cell processes of Nkx2.2 high /CNP + differentiating oligodendrocytes (Fig. 2F ,G). Altogether, these results indicate that high but transient levels of Tns3 protein characterize early immature oligodendrocytes (iOL1s and early iOL2s), being localized at their cytoplasm and cell processes (Fig. 2H,I) .

Finally, we investigated whether other Tensin family members were expressed in oligodendroglia, finding that Tns1 and Tns2 but not Tns4 were detectable at low levels in iOLs by immunofluorescence (Fig. S5A ,B), paralleling their low transcription levels compared to Tns3 ( Fig. S5C ; brainrnaseq.org). Therefore, Tns3 appears to be the main Tensin expressed during oligodendrocyte differentiation, suggesting that Tns3 function in immature oligodendrocytes is likely to be evolutionarily selected, and thus of biological importance in oligodendrogenesis.

Given the strong Tns3 expression in iOLs during postnatal myelination, we hypothesized that Tns3 expression could be enriched during remyelination in newly formed oligodendrocytes contributing to remyelination. To test this hypothesis, we performed lysolecithin (LPC) focal demyelinating lesions in the corpus callosum of adult (P90) Tns3 Tns3-V5 and wild-type mice, and assessed for Tns3 expression at the peak of oligodendrocyte differentiation (8 days postlesion) in this remyelinating model (Nait-Oumesmar et al., 1999) . We found that while nonlesioned adult brain regions contained only sparse Tns3 + iOLs (CC1 high /Olig1 cyto-high cells), remarkably many Tns3 + iOLs were detected in the remyelinating area using both V5 ( Fig.   S6A ,C, arrows) and Tns3 antibodies (Fig. S6B, arrows) . Quantification of Tns3 + cells showed a clear increase in Tns3 + iOLs around the lesion borders compared to the corpus callosum far from the lesion area (Fig. S6D) , suggesting that Tns3 expression may be a useful marker of ongoing remyelination and lesion repair. Of note, we could also detect Tns3 expression in some microglia/macrophages in the lesion area using a combination of F4/80 antibodies (Fig. S6C , arrowheads). Altogether, all these data indicate that Tns3 expression peaks at the onset of oligodendrocyte differentiation, labeling immature oligodendrocytes during both myelination and remyelination.

Tns3 knockout mice present normal numbers of oligodendroglia in the postnatal brain and still express Tns3 full length transcripts

To explore the role of Tns3 in OL differentiation, we first analyzed a Tns3 gene trap mouse line (Tns3 βgeo ) previously studied outside the CNS (Chiang et al., 2005) , where the βgeo cassette is inserted after Tns3 exon 4 ( Fig. S7A ) driving LacZ transcription and by inserting a stop poly-A sequence, predicted to be a Tns3 loss-of-function mutation. Despite the original report of postnatal growth retardation in Tns3 βgeo/βgeo mice, these mice were kept in homozygosity for several generations in C57BL/6 genetic background (Su-Hao Lo, UC Davis). We thus analyzed the impact in oligodendrogenesis in the postnatal brain of Tns3 βgeo animals. We first immunodetected βgalactosidase in OLs (Olig2 + /CC1 + and Olig2 + /PDGFRα − cells) of Tns3 βgeo postnatal brains at P21 (Fig. S7B ,C), paralleling our characterization of Tns3 expression with V5 and Tns3 antibodies. We then quantified the density of PDGFRα + OPCs or CC1 + oligodendrocytes in Tns3 βgeo/βgeo and Tns3 βgeo/+ littermates at P21, finding similar number of OPCs and oligodendrocytes in two main white matter areas (corpus callosum and fimbria; Fig.   3A -B). Moreover, quantification of three different stages of oligodendrocyte differentiation (iOL1, iOL2, and mOL) by Olig2/CC1/Olig1 immunofluorescence did not reveal changes in the rate of oligodendrocyte differentiation (proportion of each stage) in Tns3 βgeo/βgeo mice compared to control littermates ( Fig. 3C-D) . We verified the homozygosity of Tns3 βgeo allele in Tns3 βgeo/βgeo mice by PCR amplification from genomic DNA of P21 brains finding that primers recognizing Intron 4 and βgeo only produced PCR amplicons in Tns3 βgeo/βgeo mice but not when using Intron 4 flanking primers that only produced PCR amplicons in wild type mice; Fig. S7D ).

We then checked for Tns3 full-length transcripts using cDNA generated from P21 brains, and to our surprise, primers flanking Exons 17 and 31 were similarly amplified from cDNA of Tns3 βgeo/ βgeo and wild type brains (Fig. S7E ), suggesting that in the brain of Tns3 βgeo/ βgeo mice Tns3 full-length transcripts coding for Tns3 protein are still produced. Altogether, these results suggested that, unlike in other tissues (Chiang et al., 2005) , the Tns3 βgeo allele does not lead to Tns3 loss-of-function in the brain, likely through the generation of alternative spliced Tns3 variants, and is thus not suitable for assessing Tns3 function in the CNS.

We therefore generated new Tns3 knockout mouse using CRISPR/Cas9 technology, by introducing loss-of-function mutations (indels) at the beginning of Tns3 full-length coding sequence. We generated CRISPR integrative plasmids (Fig. S8A ) driving Cas9 expression and gRNAs targeting Tns3 Exon-6 at the levels of the first coding ATG, using as control, plasmids without the Tns3-targeting sequence of the gRNA. Strong cutting efficiency of two gRNAs was validated by lipofection of neural progenitors ( Fig. S8B -F). We then used these optimized tools to induce CRISPR-mediated Tns3 mutations in mouse zygotes (Methods), generating and characterizing two mouse lines having small deletions (4-deletion and 14-deletion) after the first coding ATG of Tns3 (Fig. 3E ), expected to cause frame shifts leading to Tns3 loss-offunction. Remarkably, homozygous animals were found in reduced numbers compared to mendelian ratios with many of them dying during embryonic development ( Fig. 3F ) with most homozygous animals showing major growth retardation by the second postnatal week compared with their littermates (Fig. 3G ), similar to the original report of Tns3 βgeo mice (Chiang et al., 2005) . Furthermore, we could still immunodetect Tns3 protein in CC1 + oligodendrocytes of these homozygous mice at P21 with at least two different Tns3 antibodies (Fig. 3H ,I), and detect Tns3 exons corresponding to Tns3 full-length transcript by qPCR (Fig. 3J ). Further analysis of these mice was prevented by the Covid-19 lockdown leading to the loss of these Tns3 knockout mouse lines. Altogether, these results suggest that mice carrying constitutive Tns3 loss-of-function mutations seems to escape the full Tns3 loss-of-function in the brain, by generating alternative spliced variants containing the main Tns3 full-length exons, and thus we considered these animals not suitable to study Tns3 function in oligodendrogenesis.

Finally, to assess whether TNS3 is potentially required during human development, we explore for the presence of TNS3 gene variants in the human population using the gnomAD database containing 125,748 exomes and 15,708 whole-genome sequences from unrelated individuals (Karczewski et al., 2020; Lek et al., 2016) . Homozygous predicted loss-of-function (pLoF) alleles of TNS3 were not found, and heterozygous pLoF were greatly below the expected frequency (0.1 observed/expected ratio, with 90% confidence interval of 0.05-0.19; and LOEUF of 0.19; Fig. S9A ; https://gnomad.broadinstitute.org), meaning that heterozygous loss-of-function variants of TNS3 causes ~80% developmental mortality, a rate similarly high to key neurodevelopmental genes such as SOX10 (LOEUF=0.21; Fig. S9B ), CHD7 (LOEUF=0.08; Fig. S9C ), and CHD8 (LOEUF=0.08; Fig. S9D ), contrary to less broadly required factors such as NKX2-2 (LOEUF=0.67; Fig. S9E ) and OLIG1 (LOEUF=1.08; Fig. S9F ). Therefore, TNS3 loss-offunction variants are badly tolerated in both mouse and human development.

Given the tendency of cells to escape the Tns3 loss-of-function upon constitutive knockout mutations, we decided to assess Tns3 requirement during postnatal oligodendrogenesis by inducing in vivo acute Tns3-deletion in few neural stem cells (NSCs) of the neonatal brain and tracing their cell progeny with a GFP reporter. For this, we combined the postnatal electroporation technique with CRISPR/Cas9 technology. First, we used our previously validated gRNAs targeting Tns3 at the first coding ATG (exon 6; Fig. S8 ) inserting them in an integrative CRISPR/Cas9 plasmid also expressing the GFP reporter ( Fig. 4A ), to transfect neonatal NSCs of the dorsal subventricular zone (SVZ), which generate a large number of oligodendroglial cells during the first postnatal weeks (Kessaris et al., 2006; Nakatani et al., 2013) , and focused our study on glial cells by quantifying the GFP + progeny of targeted NSCs, outside the SVZ and located in the dorsal telencephalon three weeks later (P22, Fig. 4B ). The fate of GFP + cells was determined by immunodetection of GFP and glial subtype markers (CC1 high for oligodendrocytes, PDGFRα for OPCs, and CC1 low and their unique branched morphology for astrocytes). Remarkably, brains electroporated with the CRISPR plasmids targeting Tns3 had a 2-fold reduction in GFP + oligodendrocytes compared to brains electroporated with control plasmids, while GFP + OPCs were found in similar proportions ( Fig.   4C ,C',D). The proportion of GFP + astrocytes was increased by 1.5-fold, likely as a result of the large reduction in oligodendrocytes, as the number of GFP + astrocytes was not changed (61.3 ± 10.9 in experimental versus 57.2 ± 11.8 in controls; Fig. 4C ,C',D). To assess whether the reduction in oligodendrocytes from Tns3-deleted NSCs was the consequence of a reduction in OPCs generated, we assessed for possible changes in numbers, proliferation, and survival of OPC at P11, when most cortical OPCs have not yet started differentiation. We found no differences in the proportion of GFP cells being OPCs (Fig. 4E ), nor the proliferative status of GFP + OPCs (MCM2 + /PDGFRα + cells; Fig. 4F ) between experimental and control brains, while the reduction of oligodendrocytes was already marked (Fig. 4E) , indicating that loss of Tns3 only affected the process of OPC differentiation into oligodendrocytes.

Given the expression of two Tns3 isoforms in the brain (Fig. S4F ,G), we asked whether a deletion of both isoforms would have a greater impact in oligodendrocyte differentiation.

We thus used two gRNAs efficiently cutting the beginning and the end of Tns3 coding sequence (5'-3'gRNAs, Methods), to delete the whole Tns3 locus. We found a similar reduction of oligodendrocytes in the loss of the two Tns3 isoforms than in mutations affecting only fulllength Tns3 (Fig. 4H ,H',I), suggesting that the small Tns3 isoform does not play an additional role with full-length Tns3 in oligodendrocyte formation. Altogether, these results indicate that

Tns3 loss-of-function mutations in neonatal SVZ-NSCs impair OPC differentiation without apparent changes in OPC generation and proliferation, thus suggesting that Tns3 is largely required for OPC differentiation into oligodendrocytes in the postnatal brain (Fig. 4G ).

Given the heterogeneity of CRISPR/Cas9-mediated indels and the difficulties to assess in vivo the penetrance of their Tns3 loss-of-function, to address in more detail the consequences of In order to specifically delete Tns3 in postnatal OPCs, we administered tamoxifen at P7

to Pdgfra-CreER T ; Tns3 flox/flox ; Rosa26 stop-YFP (hereafter called Tns3-iKO mice) and control pups (Pdgfra-CreER T ; Tns3 +/+ ; Rosa26 stop-YFP littermates) and analyzed its effects on oligodendrogenesis at P14 and P21 (Fig. 5A ) both in white matter (corpus callosum and fimbria) and grey matter regions (cortex and striatum). We first assessed for the efficiency of Tns3 deletion in Nkx2.2 + /GFP + iOLs from different regions by immunofluorescence using a Tns3 antibody (Sigma Ct), finding that the strong Tns3 signal present in Nkx2.2 + /GFP + iOLs of control brains was almost completely eliminated in Tns3-iKO iOLs without affecting Tns3 expression in vessels (Fig. S11B ,B',C, arrows and arrowheads versus asterisks). We then assessed for changes in oligodendrogenesis. Remarkably, the number of oligodendrocytes (CC1 + /GFP + cells) was reduced by half in all quantified regions (reduction of 38.95% in the CC, 48.60% in cortex, 50.88% in the fimbria, 38% in the striatum; Fig Altogether, these results indicate that acute deletion of Tns3 in OPCs reduces by 2-fold generation of oligodendrocytes in the postnatal brain, without major changes in OPC numbers and proliferation (Fig. 5G ).

Tensins are known to mediate integrin stabilization and activation in other cells-types (Liao and Lo, 2021) , with Tns3 been shown to bind integrin-β1 through its phosphotyrosine-binding domain and FAK through its SH2 domain in fibroblasts (Cui et al., 2004; Liao et al., 2007; Georgiadou et al., 2017) . In oligodendroglia, integrin α6β1 association with Fyn kinase is required to amplify PDGF survival signaling and to promote myelin membrane formation, by switching neuregulin signaling from a PI3K to a MAPK pathway (Colognato et al., 2004) .

Moreover, by conditional ablation of integrin-β1 in vivo, it was demonstrated that integrin-β1 signaling is involved in survival of differentiating oligodendroglia, but not required for axon ensheathment and myelination per se (Benninger et al., 2006) . We therefore, investigated the expression of genes involved in integrin signaling in the transcriptome of oligodendroglial cells. Indeed, Tns3 expression pattern in iOLs was closely matching that of Itgb1 (integrin-b1), Fyn, Bcar1/p130Cas, and Ptk2/Fak both in mouse and human oligodendroglia (Fig. S13A,B) .

Furthermore, using neural progenitor differentiation cultures, we observed co-expression of integrin-β1 and Tns3 in CNP + oligodendrocytes by immunofluorescence (Fig. S13C) , suggesting that Tns3 could relay integrin-β1-mediated survival signal in differentiating oligodendroglia. Therefore, we assessed signs of cell death in Tns3-iKO oligodendroglia by performing the TUNEL technique together with GFP and CC1 immunodetection. Interestingly we found a 5fold increase in TUNEL + cells in the dorsal telencephalon of Tns3-iKO brain, compared to control, without significant changes in non-oligodendroglial cells present in the SVZ ( Fig. 6A -C). To gain more insight into the cellular alterations and cell death of Tns3-deleted oligodendroglia, we investigated their cellular morphology and behavior by video microscopy during their differentiation in culture. To this end, we MACS-purified OPCs from Tns3-iKO and control (Tns3 flox/flox ; Rosa26 stop-YFP littermates) mice at P7, two days after administration of tamoxifen, plated them in proliferating medium for three days, and recorded their behavior during three days in the presence of differentiation medium (Fig. 6D ). Using the expression of the YFP as a readout of Cre-mediated recombination, we compared the behavior of YFP + cells (Tns3-iKO) with neighboring YFPcells (internal control) in the same cultures. In parallel, we used MACSorted cells from control mice as external control. Quantification of the proportion of YFP + cells over time showed a 20% reduction of YFP + cells (from 80% to 60%) during the 3 days in proliferation medium followed by a reduction to 50% by day 3 in differentiation medium (Fig. 6E ), suggesting possible cell death of Tns3-mutant cells. Live imaging monitoring of cell behavior showed that once YFP + cells had developed multiple branched morphology, characteristic of differentiating oligodendrocytes, they showed a 4-fold increase in their probability to die compared to YFP − cells of the same culture ( Fig. 6F -H, yellow and white arrows, respectively) or to cells from control cultures, with more pronounced cell death by the third day of culture (Fig. 6F,G) . Together, these results indicate that Tns3-iKO oligodendroglia present increased cell death both in vivo and in primary culture, at the stage when Tns3 is upregulated and cells start to developed their branched morphology, suggesting that Tns3 likely mediates β1-integrin signaling required for their survival.

To study the molecular mechanisms of Tns3 function in oligodendroglia, we first looked at P53 expression, the master transcriptional regulator of the cellular genotoxic stress response (Kastenhuber and Lowe, 2017; Aubrey et al., 2018) . Interestingly, we found a 10-fold increase in p53 + OPCs (GFP + /CC1cells) and 4-fold increase in p53 + iOLs (GFP + /CC1 + cells) in Tns3-iKO compared to control (Fig. 7A-C) , suggesting that the loss of Tns3 leads to an upregulation of p53, that together with the loss of integrin-β1 survival signal, mediates the cell death of Tns3-iKO differentiating oligodendroglia (Fig. 7D ).

The tight balance of oligodendrocyte precursor cells (OPCs) between proliferation, survival, and differentiation ensures their capacity to respond to the myelination needs of the CNS by generating new oligodendrocytes on demand, whilst avoiding the generation of brain gliomas through uncontrolled OPC proliferation. The observation that OPCs are present within demyelinating MS lesions, but fail to efficiently differentiate into myelinating cells with age and disease progression (Chang et al., 2002; Neumann et al., 2019) , together with the strong sensitivity of immature oligodendrocytes to survival/apoptotic signals (Hughes and Stockton, 2021) , suggests that efforts to foster OPC differentiation and survival of immature oligodendrocytes are a critical events for healthy aging and successful remyelination in MS patients. In this study, we combined the genome-wide binding profile of key regulators of oligodendrocyte differentiation, Olig2, Chd7, and Chd8 (Lu et al., 2000; Zhou et al., 2000; Lu et al., 2002; Zhou and Anderson, 2002; He et al., 2016; Küspert and Wegner, 2016; Zhao et al., 2018) , to identify their common gene targets, and focused our analysis on Tensin3 (Tns3), whose expression matched the onset of oligodendrocyte differentiation.

To study Tns3 expression and function, we generated several genetic tools, including CRISPR/Cas9 vectors to induce Tns3 mutations both in vivo and in vitro, a Tns3 Tns3-V5 knock-in mouse, two constitutive Tns3 knockout mice, and finally an inducible knockout (Tns3 Flox ) mouse. Using these tools, we provide several lines of evidence showing that Tns3 is upregulated in immature oligodendrocytes (iOLs) and required for normal oligodendrocyte differentiation. First, we show that Tns3 expression is strongly induced at the onset of oligodendrocyte differentiation, localized to the cytoplasm and main cell processes of iOLs, and downregulated in mature oligodendrocytes both at the transcript and protein levels, thus constituting a novel marker for iOLs, for which we provide an optimal immunofluorescence protocol with a commercial antibody (Sigma, Ct). Second, we show that during remyelination, Tns3 is also expressed in newly formed oligodendrocytes and thus could be used as a hallmark for ongoing remyelination. Third, analyzing both Tns3 βgeo gene trap mice and two Tns3 KO mice, we show that constitutive Tns3 deletion is detrimental for normal development and that the predicted loss of Tns3 full-length transcript and protein is bypassed in the oligodendroglia of surviving homozygous animals, paralleling the intolerance for TNS3 loss-of-function variants found in the human population. Fourth, in vivo CRISPR-mediated Tns3-deletion in neonatal neural stem cells from the subventricular zones leads to a 2-fold reduction of oligodendrocytes without changes in OPC generation, proliferation, and numbers. Fifth, in vivo Tns3 induced knockout (Tns3-iKO) in postnatal OPCs leads to a 2-fold reduction of differentiating oligodendrocytes without reducing the overall OPC population, both in grey and white matter brain regions. Finally, we provide evidence, by immunodetection in vivo and video microscopy of primary OPC differentiation cultures, that Tns3-iKO differentiating oligodendroglia upregulate p53, key sensor of cell stress, and present a 4-5-fold increase in apoptosis compared to control oligodendroglia, suggesting that mechanistically Tns3 function is likely required for normal oligodendrocyte differentiation at least in part by mediating integrin-β1 survival signaling in differentiating oligodendroglia.

Recent studies has started to uncover genes enriched in immature oligodendrocytes (iOLs), such as Itpr2 (Zeisel et al., 2015; Marques et al., 2016) , Enpp6 (Xiao et al., 2016) , and Bcas1 (Fard et al., 2017) , that could be used as markers for these transient cell populations, particularly interesting to label areas of active (re)myelination in the context of oligodendrocyte and myelin pathology, such as preterm brain injury and multiple sclerosis.

Here, we report for the first time that Tns3 is a hallmark of iOLs (figure 2). Tns3 is expressed at high levels in iOLs and downregulated as oligodendrocytes mature into myelinating cells, showing a complete overlap with Itpr2 transcript and protein. We found that a commercial Tns3 antibody (Millipore) also recognizes another nuclear protein that, like Tns3 in the cytoplasm, also labels at high levels iOLs, paralleling the case of CC1 antibody, which recognizes both APC and Quaking-7 proteins in oligodendrocytes (Lang et al., 2013; Bin et al., 2016) . Upon testing several antibodies, we found one (Sigma Ct) optimally labeling iOLs by immunofluorescence in brain sections and oligodendroglial cell cultures, whereas the Itpr2 commercial antibody we tried did not match this high quality iOLs immunolabeling. The use of BCAS1 antibodies has been proposed to label iOLs (Fard et al., 2017) , but has the caveat of being also expressed in mature oligodendrocytes and myelin ( Fig. S1E ; (Ishimoto et al., 2017) .

Finally, Enpp6 is very specific for iOLs at the transcript level (Xiao et al., 2016) but to our knowledge, no Enpp6-recognizing antibodies producing good quality immunodetection are yet available. Therefore, Tns3 protein expression in the CNS is a hallmark of iOLs, and the Tns3 Sigma Ct antibody is an optimal reagent to label iOLs during both myelination and remyelination.

Oligodendrocyte differentiation involves substantial generation of new membrane and cell processes composing the 40-60 myelin segments formed by mature oligodendrocytes (Hughes et al., 2018) . Actin cytoskeleton remodeling is an important driver of the OL morphological changes undergone during their differentiation (Nawaz et al., 2015; Zuchero et al., 2015) .

Tensin proteins, linking the extracellular signals received by transmembrane integrins with the actin cytoskeleton in different cell types (Liao and Lo, 2021) , are well placed to play an important role in these morphological changes. At the molecular level, it has been shown that the phosphotyrosine-binding domains of tensins interact with the NPXY motifs present in the cytoplasmic tails of integrin-β1 in a pTyr-insensitive fashion (Calderwood et al., 2003; Katz et al., 2007; McCleverty et al., 2007) , allowing tensins to bring actin filaments, through their actin binding domain, to focal adhesion sites (Liao and Lo, 2021) . Given that the extension of OL cell processes' growth cone is guided by the sequential activation of Fyn, FAK and RhoGAP (Thomason et al., 2020) , and that high levels of Tns3 protein are detected in the cell body and processes of iOLs coinciding with their large enlargement, Tns3 is thus well placed to mediate integrin signaling to the actin cytoskeleton and play an active role in this large cellular remodeling. Moreover, Integrin-β1, FAK/Ptk2, Fyn, p130Cas/Bcar1, and Tns3 are all highly expressed in iOLs (Fig. S13) . Here, using three independent approaches, we show that loss of Tns3 in iOLs reduces by half the numbers of oligodendrocytes in the postnatal brain. It is therefore very likely that Tns3 act as a mediator of integrin α6β1 signaling to promote OL survival and differentiation by mediating actin cytoskeletal remodeling. If so, exogenous activation of integrin α6β1 in cultured OPCs by Mn 2+ (Colognato et al., 2004) would not be expected to increase oligodendrogenesis in Tns3-iKO oligodendroglia. Finally, through its additional ability to bind to EGFR (Cui et al., 2004) , whose activation is another driver of oligodendroglial differentiation, Tns3 could also be required for mediation of signaling downstream of growth factor receptor activation in early iOLs; this could also explain the increased death in OLs lacking Tns3.

Programmed cell death regulates developmental oligodendrogenesis, with a large proportion of immature oligodendrocytes degenerating before the fourth week of postnatal life in mice (Barres et al., 1992; Trapp et al., 1997) . Also in the adult mouse brain, differentiating OPCs remain in the immature oligodendrocyte stage for roughly 2 days with many of them undergoing programmed cell death (Hughes et al., 2018) , indicating that this immature stage is very dependent on survival signals. Apoptotic pathways involving BCL-2 family members have been shown to regulate this oligodendroglial programmed cell death (reviewed in (Hughes and Stockton, 2021) . On study has shown that the transcription factor TFEB, involved in autophagy and lysosomal biogenesis, ensures the spatial and temporal specificity of developmental myelination by promoting the expression of ER stress genes and PUMA, a proapoptotic factor inducing Bax-Bak-dependent programmed cell death in differentiating oligodendroglia (Sun et al., 2018) . Another recent study showed that during both during homeostasis and remyelination, the activity of the primary sensor of cellular stress, nuclear factor (erythroid-derived 2)-like 2 (Nrf2), induces the expression of Gsta4, a scavenger of lipid peroxidation, with in turn controls apoptosis of immature oligodendrocytes via the mitochondria-associated Fas-Casp8-Bid-axis (Carlström et al., 2020) .

Several studies have shown that integrin-β1 signaling is required for iOL survival. Neuronalderived signals, including neuregulin and laminin-2, are received by immature oligodendrocytes through integrin-β1 signaling that would enhance the function of neuroligin as a survival factor by inducing a survival-dependence switch from the phosphatidylinositol 3kinase-Akt pathway to the mitogen-activated protein kinase (MAPK) pathway, with enhanced Benninger et al., 2006) . Also, PDGF survival signaling in OPCs and myelin formation have been shown to depend on integrin α6β1 binding to Fyn (Colognato et al, 2004) . Tensins typically reside at focal adhesions, which connect the extracellular matrix (ECM) to the cytoskeletal networks through integrins and their associated protein complexes (Kumar, 1998; Liao and Lo, 2021) , with focal adhesions mediating both outside-in and inside-out signaling pathways that regulate cellular events, such as cell attachment, migration, proliferation, apoptosis, and differentiation (Liao and Lo, 2021) . In this study, we show that Tns3 expression timing during oligodendrogenesis, parallels that of integrin-β1, and we provided evidence of Tns3 and integrin-β1 co-localization in dotted structures resembling nascent and focal adhesion in the cytoplasm and processes of immature oligodendrocytes. Therefore, given the similar apoptotic phenotypes found in differentiating oligodendroglia upon integrin-β1 (Benninger et al., 2006) or Tns3 oligodedendroglial specific knockout (this study), we suggest that Tns3 is required for integrin-β1 survival signal in immature oligodendrocytes. Moreover, we suggest that this would lead to cellular stress of Tns3-deleted differentiating oligogodendroglia and to the upregulation of p53, master regulator of cellular stress and apoptosis, which has been previously been shown to be involved in the apoptosis of human oligodendrocytes in the context of MS (Ladiwala et al., 1999; Wosik et al., 2003) and in the cuprizone demyelination mouse model (Li et al., 2008; Luo et al., 2021) .

In summary, here we have generated powerful genetic tools allowing to assess for the first time the role of Tns3 in the CNS, shown that Tns3 protein is found at high levels in the cytoplasm and main processes of immature oligodendrocytes thus constituting a new marker of this oligodendroglial stage, and demonstrated by different genetic approaches that Tns3deletion leads to a two-fold reduction in differentiating oligodendrocytes, explained at least in part by their increased apoptosis due to p53 upregulation and likely the loss of integrin-β1mediated survival signaling. Follow up studies using these tools should unravel with more detail the molecular mechanisms mediated by Tns3 not only in immature oligodendrocytes during developmental myelination but also in pathological contexts such as preterm birth dysmyelination, adult demyelination in MS and glioblastoma, this last recently associated with reduced levels of Tns3 (Chen et al., 2017) . 

CRISPOR software (http://crispor.tefor.net/) was used to design gRNAs with predicted cutting efficiency and minimal off-target and PCR amplification primers. The validation of Tns3targeting CRISPR/Cas9 system was performed in 3T3 cells by transfection with Lipofectamine 3000 of PX459 plasmids containing 4 different sgRNA sequences. After 2 days incubation, puromycin was added to medium for 4 days allowing survival of cells containing the PX459 plasmid. Three days after proliferation in fresh medium without puromycin, DNA was extracted using DNeasy blood & tissue kit (Qiagen). The target DNA for 5' Tns3 region was amplified by PCR using primers with the following sequences: Forward: 5'-AGG TGG CCT TCA GCT CAGT-3', Reverse: 5'-GCT ATC ATC CCC ACT CAC CA-3'; annealing temperature of 64°C, with the PCR product expected to be 326bp. DNA from 3' Tns3 target region was amplified using primer with the following sequences: Forward: 5'-CCA GTC AGT GGT GAC ATT GTTT-3', Reverse: 5'-ACT GTT CCC AGG TTG CTA TCAT-3'), annealing temperature of 58°C, with the PCR product expected to be 419bp.

Cutting efficiency of sgRNA was verified by T7 endonuclease I, following the beta protocol of IDTE synthetic biology for amplification of genomic DNA and detecting mutations, using PAGE. In order to generate plasmids that will insert CRISPR tools into the genome of the transfected cells and lead to permanent expression of the targeting tools, the PX458 (GFP) or PX459 (Puromycin) plasmids were subcloned into a Tol2-containing sequence backbone (obtained from Tol2-mCherry expressing plasmid kindly provided by Jean Livet, Institut de la Vision, Paris).

Tns3 V5 mice were generated at the Curie Institute mouse facility. Briefly, the Cas9 protein, the crRNA, the tracrRNA and a ssODN targeting vector for the Tns3 gene had been microinjected into a mouse egg cell, which was transplanted into a C57BL/6J-BALB/cJ female surrogate. Pups presenting HDR insertion of the V5 tag were selected after genotyping.

Tns3 KO mice were generated at the ICM mice facility. Briefly, the Cas9 protein, the crRNA, the tracrRNA and a targeting vector for the Tns3 gene had been microinjected into a mouse egg cell transplanted into a C57BL/6J female surrogate. Pups with NHEJ mutations inducing a gene frameshift were selected after genotyping and Sanger sequencing verification. Finally, only two lines containing indels of 4 and 14 nucleotide deletions were maintained and studied.

Tns3 conditional knockout mice (LoxP-Exon9-LoxP) were generated at the Transgenic Core Tamoxifen (Sigma, T5648) was dissolved in corn oil (Sigma, C-8267) and injected subcutaneously at 20mg/ml concentration at P7 (30µl) in Ctrl and Tns3-iKO animals. Brains were then collected at P21.

Postnatal brain electroporation (Boutin et al., 2008) was adapted to target the dorsal SVZ.

Briefly, postnatal day 2 (P2) pups were cryoanesthetized for 2 min on ice and 1.5 µl of plasmid was injected into their left ventricle using a glass capillary. Plasmids were injected at a concentration of 2-2.5µg/µl. Electrodes (Nepagene CUY650P10) coated with highly conductive gel (Signagel, signa250) were positioned in the dorso-ventral axis with the positive pole dorsal. Five electric pulses of 100V, 50ms pulse ON, 850ms pulse OFF were applied using a Nepagene CUY21-SC electroporator. Pups were immediately warmed up in a heating chamber and brought to their cages at the end of the experiment.

Before surgery, adult (2-3months) WT mice were weighed and an analgesic (buprenorphine, 30 mg/g) was administered to prevent postsurgical pain. The mice were anesthetized by anesthetized by induction of isoflurane (ISO-VET). Ocrygel (Tvm) was put on their eye to prevent dryness and lidocaine in cream (Anesderm 5%) was put on the ear bars to prevent pain. After cutting of the skin, a few drop of liquid lidocaine were put to prevent pain. Focal demyelinating lesions were induced by stereotaxic injection of 1µl of lysolecithin solution (LPC, Sigma, 1% in 0.9%NaCl) into the corpus callosum (CC; at coordinates: 1 mm lateral, 1.3 mm rostral to bregma, 1.7 mm deep) using a glass-capillary connected to a 10µl Hamilton syringe.

Animals were left to recover in a warm chamber before being returned into their housing cages.

Postnatal mouse were transcardially perfused with 15ml (P14) or 25ml (>P21) of 2% PFA freshly prepared from 32% PFA solution (Electron Microscopy Sciences, 50-980-495). Perfused brains were dissected out, dehydrated in 10% sucrose followed by 20% sucrose overnight, and embedded in OCT (BDH) before freezing and sectioning (16µm thickness) in a sagittal plane with a cryostat microtome (Leica).

Dissociation of cortex and corpus callosum from mice brain was done using neural tissue 

Postnatal mouse brain cryosections were dried for 20 minutes at room temperature, before adding the blocking solution (10% normal goat serum (NGS, Eurobio, CAECHVOO-OU) and 0.1% Triton X-100 in PBS) for one hour at room temperature. Primary antibodies were diluted (dilutions indicated in Table1) in the same blocking solution and incubated on the slices overnight at 4°C. After washing with 0.05% Triton X-100 in PBS, sections were incubated with secondary antibodies conjugated to AlexaFluor-488, AlexaFluor-594 and AlexaFluor-647 (Thermo, 1:1,000). Finally, cell nuclei were labeled with DAPI (1/10000, Sigma-Aldrich®, D9542-10MG), and slices are mounted in Fluoromount-G® (SouthernBiotech, Inc. 15586276).

In Situ Cell Death detection kit (Roche, 12156792910) was used to do Tunel experiment on P21 mouse brains. Briefly, tissues were processed as mentioned above with anti-GFP and anti-CC1 and fixed in fixation solution for 20min at room temperature. After washing, slices were permeabilized for 2 min in permeabilisation solution (0.1% Triton X-100 ; 0.1% sodium citrate) and TUNEL reaction mixture was put on samples for one hour at 37°C. Tissues were then mounted in Fluoromount-G®. For each Western Blot, we used 50µg of proteins denaturated for 10 minutes at 95°C with added β-mercaptoethanol (from 24X stock) and BoltTM LDS Sample Buffer (4X) (ThermoFisher, B0007). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed using precast 4-12% polyacrylamide gradient gels (ThermoFisher, NW04122BOX), submerged at 4°C in Bolt™ MOPS SDS Running Buffer (ThermoFisher, B0001) using Mini Gel Tank and Blot Module Set (ThermoFisher, NW2000). Precision Plus Protein™ All Blue protein standards (BioRad, 1610373EDU) were run alongside the samples as a protein migration control. Proteins were separated for 90 minutes at 90V, after which gels were transferred onto Amersham Protran 0.2 µm nitrocellulose membrane (Dutscher, 10600001) immersed at 4°C in NuPAGE Transfer Buffer (ThermoFisher, NP0006-1) for 90 minutes at 60V.

Following transfer, membranes were incubated for 1h in TBS-T, 10% dry milk to aid blocking of non-specific binding by the antibodies. Primary antibodies diluted in TBS-T were incubated with the membrane overnight at 4°C with shaking. After three washes in TBS-T, membranes were incubated with HRP-conjugated secondary antibodies diluted in TBS-T for 1h at 4°C with shaking, then developed using Pierce™ ECL Western Blotting Substrate (ThermoFisher, 32109) and imaged with the ChemiDoc™ Touch Imaging System (BioRad, 1708370). Western blot detection of actin was performed as loading control. Insulin (Sigma, I6634). After 3 days of proliferation, medium was replaced by growth factor depleted medium. Cell differentiation was tracked for 3 days using time-lapse video recording.

Cells were put in to a videomicroscope (Zeiss AxioObserver 7, provided by ICM-quant and CELIS facilities) with a humidified incubator at 37°C with a constant 5% CO2 supply. Images for both FITC and bright field were acquired every 10 minutes.

ChIP-seq assays were performed as described previously , using iDeal ChIPseq kit for Transcription Factors (Diagenode, C01010055). Briefly, O4 + MACSorted cells were fixed in 1% formaldehyde (EMS, 15714) for 10 min at room temperature and the reaction was quenched with 125 mM glycine for 5 min at room temperature. Lysates were sonicated with a Bioruptor Pico sonicator (Diagenode, total time 8 min) and 4μg of antibodies were added to sheared chromatin (from 4 million cells for Olig2 and from 1 million cells for histone marks) and incubated at 4°C overnight on 10rpm rotation. Antibodies used were: mouse anti-Olig2 antibody (Millipore, MABN50), rabbit anti-H3K4me3 antibody (Active motif, 39060), rabbit anti-H3K27Ac antibody (Active motif, 39034), rabbit anti-H3K4me1 antibody (Ozyme, 5326T), mouse anti-H3K27me3 antibody (Abcam, ab6002). Mock (Rabbit IgG) was used as negative control. Chromatin-protein complexes were immunoprecipitated with protein A/G magnetic beads and washed sequentially according to the manufacturer (Diagenode, C01010055). DNA fragments were then purified using IPure beads v2 (Diagenode, C01010055). Input (nonimmunoprecipitated chromatin) was used as control in each individual experiment.

The ChIP-seq libraries were prepared using ILLUMINA Truseq ChIP preparation kit and sequenced with ILLUMINA Nextseq 500 platform.

All ChIP-seq analysis were done using the Galaxy Project (https://usegalaxy.org/). Reads were trimmed using Cutadapt (--max-n 4) and Trimmomatic (TRAILING 1; SLIDINGWINDOW 4 and cutoff 20; LEADING 20; MINLEN 50), and mapped using Bowtie2 onto mm10 mouse reference genome (-X 600; -k 2; --sensitive). PCR-derived duplicates were removed using PICARD MarkDuplicates. Bigwig files were generated with bamCoverage (binsize=1). Peak calling was performed using MACS2 callpeak with Input as control and with options: --qvalue 0.05; -nomodel; --keep-dup 1; --broad (only for histone marks). Blacklisted regions were then removed using bedtools Intersect intervals.

Visualization of coverage and peaks was done using IGV ((Robinson et al., 2011);  http://software.broadinstitute.org/software/igv/home). Intersection and analysis of bound genes were done using Genomatix (https://www.genomatix.de/). Chd7, Chd8 and Mock ChIPseq datasets are from . Heatmap was done using R (4.0) using pheatmap package. GO analysis was done using Enrichr GO Biological Process 2021.

Two replicates were done for Olig2, with one of them of better quality (53,960 peaks for replicate 1 and 14,242 peaks for replicate 2). Only the peaks found in both replicates (6,781) and the peaks from replicate 1 which were found in regulatory elements (13948) were considered (16578 in total). Three replicates were done for H3K4me3, two replicates were done for H3K27me3 and one replicate was done for H3K27Ac and H3K4me1. Intersection of these datasets was done using bedtools Intersect intervals.

Peaks overlapping with regions between 1000bp upstream of transcription start site (TSS) and 10bp downstream of TSS were identified as "promoters" (Genomatix). "Active promoters" were represented by peaks for H3K4me3 and H3K27Ac. "Repressed promoters" were represented by peaks for H3K27me3 and no active marks. "Poised promoters" were represented by peaks for H3K4me1 and no active or repressed mark. Regions outside promoters containing histone marks were considered as "enhancers". "Active enhancers" were represented by peaks for H3K27Ac. "Repressed enhancers" were represented by peaks for H3K27me3 and no active marks. "Poised enhancers" were represented by peaks for H3K4me1 and no active or repressed mark. Genes were considered associated if the peaks were present in the promoter or within a range of 100kb from the middle of the promoter and the gene expression was medium to high ("active"), low ("poised") or not ("repressed") expressed (based on control RNA-seq dataset in GSE116601).

Raw data were downloaded from GEO datasets GSE107919 and GSE116601 and processed throught the Galaxy Project (https://usegalaxy.org/) using RNAstar for alignment on mm10 reference genome and featureCounts to obtain counts. CPM (count per million), FPKM and statistical analysis were done with R (4.0) using edgeR package.

Using control RNA-seq dataset in GSE116601, genes were classified as not (below first quartile), low (between first quartile and mean), medium (between mean and third quartile) and high (above third quartile).

Counts per gene were downloaded from GEO datasets GSE75330 and GSE95194, and processed in R (4.0) using the following packages: Seurat (3.0) for data processing, sctransform for normalization, and ggplot2 for graphical plots. Seurat objects were first generated for each dataset independently using CreateSeuratObject function (min.cells = 5, min.features = 100). Cell neighbors and clusters were found using FindNeighbors (dims = 1:30) and FindClusters (resolution = 0.4) functions. Clusters were manually annotated based on the top 50 markers obtained by the FindAllMarkers, adopting mainly the nomenclature from Marques 2016. Using the subset function, we selected only the clusters containing neural progenitors and oligodendroglia cells. Using the merge function, we combined both oligodendroglial datasets into a single Seurat object (OLgliaDevP) containing 5516 cells. The new object was subjected to NormalizeData, FindVariableFeatures, ScaleData, RunPCA, and RunUMAP functions with default parameters. Different OPC clusters were fused into a single one keeping apart the cycling OPC cluster. For DimPlots and Dotplots, clusters were ordered by stages of oligodendrogenesis from neural stem cells (NSCs) to myelinating OLs. R script is provided as supplementary file.

Raw data files have been deposited in the NCBI Gene Expression Omnibus under accession number GEO: XX

Further information and requests for reagents may be directed to, and will be fulfilled by the corresponding author Carlos Parras (carlos.parras@icm-institute.org).

We thank D. Bergles for the PDGFRα::CreER T mice. All animal work was conducted at the ICM PHENOPARC Core Facility. Data generated relied on ICM Core Facilities: PHENO ICMice, iGenSeq, iVector, CELIS, Histomics, and ICM Quant, and we thank all personnel involved for their contribution and help. The Core Facilities were supported by the "Investissements 

GnomAD data for TNS3 and key regulators of oligodendrogenesis (SOX10, CHD7, CHD8, NKX2-1, and OLIG1) showing frequency and scores of different genetic variants in the human population: synonymous, missense, and pLOF (nonsense, splice acceptor, and splice donor variants). The numbers highlighted colored squares correspond to the LOEF (loss-of-function observed/expected upper bound fraction), that it is a conservative estimate of the observed/expected ratio. Low LOEUF scores indicate strong selection against predicted lossof-function (pLoF) variation in a given gene, while high LOEUF scores suggest a relatively higher tolerance to inactivation. Note that TNS3, like SOX10, CHD7, and CHD8 have very low LOEUF scores indicating high intolerance for their inactivation, contrary to NKX2-1 and OLIG1, which are show more tolerance to their inactivation. 

How does p53 induce apoptosis and how does this relate to p53-mediated tumour suppression?

Cell death and control of cell survival in the oligodendrocyte lineage

) β1-Integrin Signaling Mediates Premyelinating Oligodendrocyte Survival But Is Not Required for CNS Myelination and Remyelination

Oligodendrocyte Development and Plasticity. Cold Spring Harbor Perspectives in Biology 8

The oligodendrocyte-specific antibody 'CC1' binds Quaking 7

Efficient In Vivo Electroporation of the Postnatal Rodent Forebrain

Genetic identification of cell types underlying brain complex traits yields insights into the etiology of Parkinson's disease

A phosphorylation switch controls the spatiotemporal activation of Rho GTPases in directional cell migration

Gsta4 controls apoptosis of differentiating adult oligodendrocytes during homeostasis and remyelination via the mitochondriaassociated Fas-Casp8-Bid-axis

Hominin-specific regulatory elements selectively emerged in oligodendrocytes and are disrupted in autism patients

Single-cell transcriptomic reveals molecular diversity and developmental heterogeneity of human stem cell-derived oligodendrocyte lineage cells

Premyelinating oligodendrocytes in chronic lesions of multiple sclerosis

Musashi-1 Enhances Glioblastoma Cell Migration and Cytoskeletal Dynamics through Translational Inhibition of Tensin3

Inactivation of tensin3 in mice results in growth retardation and postnatal lethality

Integrins direct Src family kinases to regulate distinct phases of oligodendrocyte development

CNS integrins switch growth factor signalling to promote target-dependent survival

Multiple sclerosis

Epidermal growth factor modulates tyrosine phosphorylation of a novel tensin family member, tensin3

Chromatin remodeler Ep400 ensures oligodendrocyte survival and is required for myelination in the vertebrate central nervous system

Transcriptional and Epigenetic Regulation of Oligodendrocyte Development and Myelination in the Central Nervous System

BCAS1 expression defines a population of early myelinating oligodendrocytes in multiple sclerosis lesions

The oligodendrocyte-type-2 astrocyte cell lineage is specialized for myelination

Glycolytic oligodendrocytes maintain myelin and long-term axonal integrity

AMPK negatively regulates tensin-dependent integrin activity

A single-cell atlas of entorhinal cortex from individuals with Alzheimer's disease reveals cell-type-specific gene expression regulation

Chd7 cooperates with Sox10 and regulates the onset of CNS myelination and remyelination. Nat Neurosci advance online publication

Premyelinating Oligodendrocytes: Mechanisms Underlying Cell Survival and Integration

Myelin remodeling through experience-dependent oligodendrogenesis in the adult somatosensory cortex

Mice lacking BCAS1, a novel myelin-associated protein, display hypomyelination, schizophrenia-like abnormal behaviors, and upregulation of inflammatory genes in the brain

NG2+ CNS Glial Progenitors Remain Committed to the Oligodendrocyte Lineage in Postnatal Life and following Neurodegeneration

Putting p53 in Context

A reciprocal tensin-3-cten switch mediates EGF-driven mammary cell migration

Competing waves of oligodendrocytes in the forebrain and postnatal elimination of an embryonic lineage

Signaling by integrin receptors

SomethiNG 2 talk about-Transcriptional regulation in embryonic and adult oligodendrocyte precursors

Induction by Tumor Necrosis Factor-α and Involvement of p53 in Cell Death of Human Oligodendrocytes

Adenomatous Polyposis Coli Regulates Oligodendroglial Development

Oligodendroglia metabolically support axons and contribute to neurodegeneration

Inhibition of p53 transcriptional activity: a potential target for future development of therapeutic strategies for primary demyelination

Tensins -emerging insights into their domain functions, biological roles and disease relevance

The phosphotyrosine-independent interaction of DLC-1 and the SH2 domain of cten regulates focal adhesion localization and growth suppression activity of DLC-1

Mosaic Analysis with Double Markers Reveals Tumor Cell of Origin in Glioma

Common developmental requirement for Olig function indicates a motor neuron/oligodendrocyte connection

Sonic hedgehog--regulated oligodendrocyte lineage genes encoding bHLH proteins in the mammalian central nervous system

Differential Role of p53 in Oligodendrocyte Survival in Response to Various Stresses: Experimental Autoimmune Encephalomyelitis, Cuprizone Intoxication or White Matter Stroke

Oligodendrocyte precursor survival and differentiation requires chromatin remodeling by Chd7 and Chd8

Oligodendrocyte heterogeneity in the mouse juvenile and adult central nervous system

Tensin3 Is a Negative Regulator of Cell Migration and All Four Tensin Family Members Are Downregulated in Human Kidney Cancer

Motor skill learning requires active central myelination

Oligodendrocytes in the Mouse Corpus Callosum Maintain Axonal Function by Delivery of Glucose

Wrapped to Adapt: Experience-Dependent Myelination

Progenitor cells of the adult mouse subventricular zone proliferate, migrate and differentiate into oligodendrocytes after demyelination

Ascl1/Mash1 Promotes Brain Oligodendrogenesis during Myelination and Remyelination

Actin Filament Turnover Drives Leading Edge Growth during Myelin Sheath Formation in the Central Nervous System

Remyelination and ageing: Reversing the ravages of time

Distinct Distribution of the Tensin Family in the Mouse Kidney and Small Intestine

Brain cell type-specific enhancer-promoter interactome maps and disease-risk association

Preservation of a remote fear memory requires new myelin formation

Chromatin remodelers in oligodendroglia

A myelin-related transcriptomic profile is shared by Pitt-Hopkins syndrome models and human autism spectrum disorder

Control of oligodendrocyte differentiation by the Nkx2.2 homeodomain transcription factor

Integrative genomics viewer

Disruption of Oligodendrogenesis Impairs Memory Consolidation in Adult Mice

Terminal differentiation of myelin-forming oligodendrocytes depends on the transcription factor Sox10

Spatiotemporal Control of CNS Myelination by Oligodendrocyte Programmed Cell Death through the TFEB-PUMA Axis

Multiple cell populations in the early postnatal subventricular zone take distinct migratory pathways: a dynamic study of glial and neuronal progenitor migration

Fuss B (2020) The oligodendrocyte growth cone and its actin cytoskeleton: A fundamental element for progenitor cell migration and CNS myelination

Differentiation and Death of Premyelinating Oligodendrocytes in Developing Rodent Brain

Myelin degeneration and diminished myelin renewal contribute to age-related deficits in memory

Extracellular cues influencing oligodendrocyte differentiation and (re)myelination

Oligodendrocyte injury in multiple sclerosis: a role for p53

Rapid production of new oligodendrocytes is required in the earliest stages of motor-skill learning

Myelin plasticity: sculpting circuits in learning and memory

Neuron class-specific responses govern adaptive myelin remodeling in the neocortex

Olig2 Targets Chromatin Remodelers to Enhancers to Initiate Oligodendrocyte Differentiation

Pioneer transcription factors, chromatin dynamics, and cell fate control

Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq

Dual Requirement of CHD8 for Chromatin Landscape Establishment and Histone Methyltransferase Recruitment to Promote CNS Myelination and Repair

The bHLH transcription factors OLIG2 and OLIG1 couple neuronal and glial subtype specification

Identification of a Novel Family of Oligodendrocyte Lineage-Specific Basic Helix-Loop-Helix Transcription Factors