Functional genomic analyses highlight a shift in Gpr17‐regulated cellular processes in oligodendrocyte progenitor cells and underlying myelin dysregulation in the aged mouse cerebrum

Abstract Brain ageing is characterised by a decline in neuronal function and associated cognitive deficits. There is increasing evidence that myelin disruption is an important factor that contributes to the age‐related loss of brain plasticity and repair responses. In the brain, myelin is produced by oligodendrocytes, which are generated throughout life by oligodendrocyte progenitor cells (OPCs). Currently, a leading hypothesis points to ageing as a major reason for the ultimate breakdown of remyelination in Multiple Sclerosis (MS). However, an incomplete understanding of the cellular and molecular processes underlying brain ageing hinders the development of regenerative strategies. Here, our combined systems biology and neurobiological approach demonstrate that oligodendroglial and myelin genes are amongst the most altered in the ageing mouse cerebrum. This was underscored by the identification of causal links between signalling pathways and their downstream transcriptional networks that define oligodendroglial disruption in ageing. The results highlighted that the G‐protein coupled receptor Gpr17 is central to the disruption of OPCs in ageing and this was confirmed by genetic fate‐mapping and cellular analyses. Finally, we used systems biology strategies to identify therapeutic agents that rejuvenate OPCs and restore myelination in age‐related neuropathological contexts.


| INTRODUC TI ON
Ageing in the brain is accompanied by a gradual decline in neuronal networking and synaptic plasticity which are needed for learning and cognitive function. Notably, neuronal numbers are largely unaltered in the ageing human brain (Fabricius et al., 2013;Pelvig et al., 2008). In comparison, there is evidence of gradual losses in oligodendrocytes and myelin in ageing and that these changes are key factors in cognitive decline and to decreased capacity for repair following pathology (Vanzulli et al., 2020). Sustaining myelin and oligodendrocytes throughout life is, therefore, critical and is the function of a reservoir of oligodendrocyte progenitor cells (OPCs) . The underlying causes of myelin loss in ageing are unresolved, but there is increasing evidence that a major factor may be the decline in OPC regenerative capacity (Azim et al., 2017;Neumann et al., 2019). Hence, unravelling the fundamental changes in the ageing brain is a key strategy for developing new approaches to promote repair in neurodegenerative dis-

eases, including Multiple Sclerosis (MS) and Alzheimer's disease (AD).
Transcriptomic studies have become increasingly important in understanding ageing processes in human and rodent oligodendrocytes (Azim et al., 2017(Azim et al., , 2020de la Fuente et al., 2020;Jäkel et al., 2019;Marques et al., 2016;Soreq et al., 2017). Here, using a combined transcriptomic and neurobiology approach we have identified essential oligodendroglial genes amongst the most dysregulated in the ageing mouse cerebrum, most notably Gpr17, which specifically decorates a subpopulation of differentiation committed OPCs (COPs) that are in transition to mature myelinating oligodendrocytes (MOLs) and react rapidly to brain pathology . In addition, we determined specific cellular signalling pathways and transcriptional networks that characterise aged oligodendroglia. Finally, we used novel in silico pharmacogenomics strategies for the identification of therapeutic agents that stimulate the transcriptional networks for driving the regeneration of OPCs following demyelination and have therapeutic potential in MS and neurodegenerative diseases.

| RNA-seqtranscriptomeoftheageingmouse cerebrum
The most prominent age-related changes in the brain were explored by generating RNA-seq profiles of dissected brain cerebrum from 1-month-old adult and 18-month aged mice (Figure 1a-c) and further F I G U R E 1 Transcriptomic characterisation of ageing-induced genes in the brain. (a) QC of Datasets, analysis and dispersion plot of normalised mean gene counts. (b) MA plot illustrating the differential expression analysis and identified 1706 genes significantly altered between the two groups (FDR <0.01 or pADJ <0.01) using DEseq2 (V.1.4.2). (c) Heatmap of the most altered genes in the ageing cerebrum ranked by FDR values and colour intensity relative to log 2 fold change. (d) Major ageing-induced gene changes (threshold genes at FDR <0.05) represented by GO analysis revealing Extracellular Matrix (ECM) Organisation, Gliogenesis, Neurogenesis and Myelination among the most altered Biological Pathways. (e) Network analysis of predicted protein-protein interaction performed with STRING (V.10.5) identified an alteration of the major processes and highlighting Cell Cycle (Red, FDR<0.0127), Cell Differentiation (Green, FDR<0.0307) and Inflammatory Response (Blue, FDR<0.0243, PPI Enrichment p-Value <1.0e-16) investigating altered signalling and transcriptional networks using pathway analysis (ConsensusPathDB), functional protein-protein (STRING V10.5) interactions (Figure 1d, e) (Herwig et al., 2016;Szklarczyk et al., 2015), and protein-chemical (STITCH v5.0) network analysis (Kuhn et al., 2008). A key finding was the predominance of oligodendroglial genes amongst the most significantly altered genes in the whole brain ( Figure 1c). The most temporally regulated gene was Gpr17, which in the brain is expressed exclusively in a subset of rapidly reacting oligodendroglial cells, specifically in an intermediate stage between OPCs and terminally differentiated myelinating oligodendrocytes (MOLs) (Viganò et al., 2016). Single-cell RNA-seq of oligodendrocyte lineage cells (Marques et al., 2016) has identified the expression of Gpr17 in multiple clusters that can be collectively defined as 'differentiation committed OPC' (COPs) ( Figure   S1). In addition, the highest-ranked genes altered in ageing were the major myelin-related genes, Mog, Plp1, Cnp and Ugt8a, as well as the less well-known myelin proteins Cldn11 (Bronstein et al., 2000), Tspan2 (Yaseen et al., 2017), and Mal, which regulates recruitment of PLP in myelin (Bijlard et al., 2016). These trends were verified by Gene Ontology (GO) analysis, which identified the main biological processes as those associated with Extracellular Matrix (ECM) Organisation and Gliogenesis/Differentiation, and specifically oligodendrocyte differentiation and myelination ( Figure 1d). STRING Network Visualisation revealed that the most transcriptionally reshaped landscapes in the ageing cerebrum were associated with the control of cell cycle, and protein sub-networks coupled to ECM remodelling and myelination, together with a transcriptional subnetwork associated with Gpr17 ( Figure 1e). The ECM plays a pivotal role in oligodendrocyte differentiation (Lourenço & Grãos, 2016) and increased stiffness of the ECM is related to age-related deterioration of OPC function . Overall, these unbiased statistical analyses signify oligodendroglial genes as highly susceptible to age-related changes in the mouse cortex.

| AgedOPCtranscriptionalsignatureand myelination transcriptional networks
To provide insight into the stage-specific transcriptional signatures of aged OPCs ( Figure 2) and MOLs ( Figure S2), we performed a metaanalysis of our RNA-seq database against published datasets (Zhang et al., 2014). The results confirmed the most altered processes in aged MOLs were associated with myelination ( Figure S2a,b), and at the core was Egfr (epidermal growth factor receptor) ( Figure S2c), which has recognised importance in oligodendrocyte regeneration and myelin repair (Aguirre et al., 2007); interestingly, our analysis implicates dysregulation of a novel Egfr-Vinculin-Gelsolin-Cldn11 axis in the age-related changes in myelination ( Figure S2d), whereby the mechanosensitive function of EGFR is transduced by Vcl (Vinculin) and Gsn (Gelsolin) which, with Cldn11 (claudin-11), regulate the anchoring of the actin cytoskeleton to the ECM through integrins, that are essential for myelination (Bronstein et al., 2000;Zuchero et al., 2015). In aged OPCs, GO analysis demonstrated the highest-ranked shifts in the cellular machinery were related to neural cell development, negative regulation of cell signalling and ECM organisation ( Figure 2a). Functional Protein Interaction Network Analysis (STRING) uncovered the key aged OPC gene networks with the largest transcriptomic hub were related to the cell cycle operating downstream of signalling via the ECM and a Pdgfra-Gpr17 axis ( Figure 2c). Further exploration of age-induced OPC gene networks unravelled Gpr17 as a multifactorial regulator, central to numerous pro-oligodendroglial mechanisms, in addition to its known receptor function (uracil nucleotides and cysteinyl leukotrienes) in COPs, during the transition between OPCs and MOLs (Chen et al., 2009). Our analysis identified novel interactions between Gpr17 and OPC differentiation, synaptic signalling and the ECM, together with prominent interactions between Gpr17 and other G-protein couple receptors, including P2 yr12, which mediates OPC-ECM interactions that regulate differentiation (Dennis et al., 2012). Gpr17 is an upstream hub for genes that encode for synaptic proteins in OPCs (Figure 2d), via the cell-adhesion protein Dchs1 (Dachsous Cadherin-Related 1) and Rasgrf1 (Ras Protein Specific Guanine Nucleotide Releasing Factor 1), which play essential roles in synaptic plasticity (Miller et al., 2013;Seong et al., 2015), and in the aged OPC gene network interconnect Gpr17 with Cacng4 (or Stargazin), together with the synaptic proteins Shank3, Homer2, Nrxn1/2 and Nlgn3, which regulate synaptic targeting of AMPA receptors and bidirectional stabilisation of the pre-and post-synaptic membranes (L. Chen et al., 2000;Dean & Dresbach, 2006;Shiraishi-Yamaguchi & Furuichi, 2007). Notably, Stargazin targets AMPA receptors to the OPC cell membrane (Zonouzi et al., 2011), and AMPA receptors regulate OPC proliferation, differentiation and myelination (Larson et al., 2016). The aged OPC transcriptional signature locates Gpr17 at the core of these OPC signalling networks that are most altered in the ageing brain.

| DysregulationofGpr17andoligodendrocyte differentiation in ageing
To investigate how ageing OPC regulatory networks are translated into cellular changes, we examined substages of the OL lineage in the Corpus Callosum (CC) in Pdgfra-CreERT2:Rosa26R-YFP and Gpr17-iCreERT2xCAG-eGFP mice ( Figure 3). First, Pdgfra-CreERT2:Rosa26R-YFP mice aged 3-and 18-months were injected with tamoxifen twice a day for 5 days to induce YFP expression in OPCs. After 10 days following genetic recombination, immunostaining was performed for NG2, Gpr17 and APC to identify the key stages between OPCs, COPs and terminally differentiated MOLs

| Identificationofsmallmoleculesto rejuvenate aged OPCs
Both the transcriptomic and fate-mapping/immunohistochemical findings demonstrate that disruption of OPCs and MOLs are major factors in the ageing brain. We used the SPIED/CMAP database to identify small molecules that recapitulate transcriptional changes in younger OPCs (Figure 4), as previously described (Azim et al., 2017;Rivera & Butt, 2019). We interrogated the gene sets for young adult and aged OPCs generated above, together with previously curated single-cell RNA-sequencing gene sets of young adult OPCs. In this strategy, young adult OPC-core genes were transformed to coexpression hub genes against drug connectivity mapping databases (Williams, 2013), that highlight master regulators and identified F I G U R E 2 Age-related transcriptional network alterations in OPCs. (a,b) GO analysis of OPC Biological Processes altered in ageing (a), from 322 core genes identified by meta-analysis of RNA-seq database of OPCs (b). (c) STRING analysis of the predicted interactions of OPC genes altered in ageing (PPI Enrichment p-value <1.0e-16); the circles represent groups of genes active along common pathways. (d) Highlighted Gpr17 node and key interactions with Cell Cycle, Synapses and ECM nodes F I G U R E 3 Dysregulation of Gpr17 and oligodendrocyte differentiation in the mouse cerebrum. (a) Immunostaining of Pdgfra-CreERT2:Rosa26R-YFP mice aged 3-(top panels) and 18-months (lower panels) 10 days after genetic recombination, demonstrating a reduction in the number of NG2+ OPCs (left-hand panels), Gpr17+ COPs (middle panels), and APC+terminally differentiated MOLs (righthand panels) in the cerebrum. Scale bars = 50 µm. (b) Quantification of total cell numbers per constant cerebral FOV showing a dramatic loss of NG2+ OPCs and Gpr17+ COPs, together with a significant decrease in APC+MOLs (n = 3 mice per group; *p < 0.05, ***p < 0.001, unpaired t tests). (c, d) Fate-mapping of Pdgfra-YFP+OPCs in defined differentiation stages, expressed as total number of cells per constant cerebral FOV (c) and as a proportion of the total number of YFP+cells (d), illustrating a marked decline in OPC differentiation into Gpr17+ COPs and a complete loss of cells differentiating into APC+MOLs over this period (n = 3 mice per group; *p < 0.05, **p < 0.01, unpaired t tests). (e) Chromogenic characterisation of Gpr17 expression in 1-and 18-month-old cerebrum in wild-type C57BL/6 mice, indicating a loss of Gpr17+ expression in ageing; Gpr17 densely decorates somata and processes at 1-month, whereas at 18-month Gpr17+ COPs are either dimly immunostained (black arrows), or in many cases, only the cell somata are immunostained (white arrows); scale bars = 50 µm. (f) qPCR quantification of Gpr17 expression in young and aged cerebrum; data are expressed as 2-dCt (n = 3 mice per group; *p < 0.05, unpaired t-test). (g) Fate-mapping of Gpr17+/GFP+COPs immunolabelled with NG2 for all OPCs in the cerebrum of Gpr17-iCreERT2xCAG-eGFP mice 10 days after recombination; immunostaining shows a gradual age-related decline in the density of OPCs and COPs (scale bars = 50 µm in main panels and 25 µm in insets) Gpr17 as the most highly correlated hub (Figure 4a), which fully validates the genomic and neurobiological data presented above.
We then used two distinct approaches to identify small molecules that have the potential to rejuvenate aged OPCs, by interrogating the core OPC genes across the entire SPIED/CMAP database, presented in Figure 4b as a dimensionality reduction plot, in addition to a STRING chemical-protein target analysis of all OPC genes differentially expressed in the ageing brain ( Figure 4c). Significantly, these two separate approaches identified the same small molecules with the potential to specifically rejuvenate 'stemness' in aged OPCs (Figure 4b,c), and none of these small molecules were predicted to act on MOLs ( Figure S3). In OPCs, a number of cardiac glycosides (digoxin, digoxigenin and ouabain) were highlighted as having OPC rejuvenation potential by regulation of mTOR signalling (Figure 4b,c). The small molecule with the highest number of associated target genes was LY294002, which was at the centre of the OPC rejuvenating drug network (Figure 4b,c), and is a known modulator of PI3 K-Akt-mTOR signalling, a key regulator of OPC differentiation and myelination (Ishii et al., 2019). Analysis of the biological processes of LY294002 target genes (TGs) in OPCs, using the Enrichr webtool (see Experimental Procedures), identified positive regulation of transcription, as well as ECM interactions and regulation of cell proliferation, as key mechanisms of action LY294002 in OPCs ( Figure 4d). These analyses highlighted LY294002 as a potential therapeutic strategy for rejuvenating OPCs in the ageing brain.

| ThesmallmoleculeLY294002identifiedbyin silico pharmacogenomics promotes oligodendrocyte regeneration and remyelination in older mice
To assess the effects of LY294002 on the capacity of adult OPCs to regenerate MOLs, we analysed the remyelination power of OPCs in vivo following intraventricular infusion of the demyelinating agent lysolecithin (LPC) in Sox10-EGFP mice, which identifies oligodendroglial cells at all stages of differentiation (Azim & Butt, 2011).
LPC (2%), or sterile vehicle in controls, were administered by intraventricular injection in mice aged on average 6 months, which was selected because this age is a point of inflection, after which there is a decline in the rate and overall extent of remyelination, owing to diminished OPC regenerative capacity (Crawford et al., 2016). At 5 days post-injection (DPI), LY294002 was administered by osmotic F I G U R E 4 Pharmacogenomic identification of therapeutic agents for rejuvenating OPCs in ageing contexts. (a) Overview of the metaanalyses performed for assembling transcriptional signatures using datasets generated in the present study and singlecell datasets of young OPCs for the detection of master regulators; combined gene lists were interrogated via the CMAP database. (b) Visualisation of obtained small molecules in a principle component plot by their target genes (TGs) reflected in the size of points and coloured using Pearson's correlation scores. (c) STITCH protein target analysis of pharmacogenomically-derived small molecules predicted to rejuvenate OPCs, highlighting LY294002 operating via the mTOR pathway (Red, FDR<9.3e-06) (PPI enrichment p-value 3.11e-16). (d) Biological Processes of LY294002 target genes (TGs) using the Enrichr webtool pump to provide a final concentration of 2 µM in the CSF, calculating for the dilution effect in the ventricular volume (Azim & Butt, 2011). The cell proliferation marker EdU was administered at 5 and 6 DPI, for fate-mapping of newly formed OLs (NFOLs), and brains were analysed at 10 and 14 DPI (Figure 5a). Immunolabelling for MBP confirmed evident demyelination in the CC at 14 DPI following LPC injection, together with an apparent decrease in the overall number of Sox10+ oligodendroglia (OPCs and MOLs) and APC+ MOLs, compared to controls, and these were evidently improved by LY294002 treatment (Figure 5b). These effects of LY294002 were confirmed by qPCR of microdissected CC from mice treated with LPC or LPC+LY294002, compared to age-matched untreated mice ( Figure 5c). The results demonstrate that the major OPC and MOL transcripts Pdgfra, Plp1 and MBP were all increased significantly in LY294002, compared to LPC and untreated controls (Figure 5c). In addition, LY294002 had pro-oligodendroglial and anti-inflammatory effects compared to LPC treatment, for example, Igf1 and Bmp4 are increased, whereas Lif and Stat1 are decreased (Figure 5c). Finally, we analysed the changes in distinct oligodendrocyte differentiation stages by immunolabelling for the pan-oligodendroglial marker Olig2, together with EdU and APC (Figure 5d), to identify and quantify total numbers of Olig2+/APC-OPCs and Olig2+/APC+ MOLs (Figure 5d,g,h). In addition, MBP immunostaining in MOLs is restricted to the myelin sheaths, whereas MBP is also expressed in the somata of NFOLs (Figure 5e,f, arrows), and the former were prevalent in demyelinated lesions following treatment with LY294002 ( Figure 5i). To assess more precisely the level of remyelination, we used the myelin index (MI) which is a measure of the numerical density of myelin sheaths that cross the z-plane in the CC (Figure 5j), as previously detailed . The data demonstrate that at 14DPI, compared to controls, there was a significant increase in (e,f) Higher magnification to illustrate non-myelin forming oligodendrocytes (NFOLs) and myelin-forming oligodendrocytes (MYOLs) that had incorporated EdU earlier in the lineage. (g-j) Quantification of Sox10-EGFP+/APC-cells and PDGFRα+ OPCs in the CC (g), total numbers of APC+OLs (h), mature myelinforming MYOLs (i), and the myelin index (i); ns = no significance; *=p < 0.05; **=p < 0.01; ***=p < 0.001; unpaired students t tests LY294002 validate the in silico pharmacogenomic analysis that identifies multiple small molecules with the potential to rejuvenate OPC stemness and promote remyelination and repair.

| DISCUSS ION
Age-related changes in myelination are proposed to be a major factor in cognitive decline and are implicated in myelin loss in AD and the ultimate failure of remyelination and repair in MS, although the underlying mechanisms remain unclear Vanzulli et al., 2020). Our differential transcriptomic analysis demonstrates that oligodendroglial genes are amongst the most significantly dysregulated in the mouse cerebrum in natural ageing.
Notably, our results highlight Gpr17 as a major factor affected during oligodendrocyte degeneration in the ageing brain. Moreover, we unravelled key transcriptional networks and signalling pathways that are central to age-related dysregulation of myelin turnover. Finally, we identified specific pro-oligodendroglial small molecules that rejuvenate OPC stemness and promote remyelination and repair. This study unravels new mechanisms in natural ageing and in neurodegenerative diseases.

| Oligodendroglialtranscriptomicnetworksare significantly altered in ageing
Transcriptomic analysis identified dysregulation of multiple biological processes that are critical for normal brain function, most sig- repair (Aguirre et al., 2007), and our findings indicate Egfr signalling is pivotal to multiple transcriptional networks and signalling pathways that regulate age-related changes in oligodendrocytes.

| DisruptionofGpr17+OPCsinageing
Changes in OPCs were a major hallmark of the ageing brain and specifically dysregulation of Gpr17. In the adult healthy brain, OPCs normally proliferate at very low levels, but mediate rapid repair responses to injury with increased proliferation and differentiation to mature myelinating cells (Psachoulia et al., 2009). Interestingly, Gpr17 specifically decorates a subset of OPCs responsible for such rapid reaction to damage (Viganò et al., 2016).

| DysregulationofGpr17expressionin aged OPCs
The transcriptomic and cell biological data all point to Gpr17 as being central to age-related changes in the brain. As mentioned above, Gpr17 is specifically expressed by a subset of OPCs that are normally quiescent but rapidly react to insults such as brain ischaemia (Viganò et al., 2016), suggesting that they may serve as a 'reservoir' of cells specifically devoted to repair purposes. Although these cells fail to repair damage in excessive inflammatory milieu , under 'permissive conditions' (i.e., in the presence of low inflammation levels), they generate myelinating oligodendrocytes and ameliorate damage (Coppolino et al., 2018). In line with data showing that increased inflammation in the aged brain is associated with reduced repair abilities, our data indicate there is a marked reduction of NG2+ OPCs and MOL at 18 months, consistent with a recent study indicating aged OPCs lose their stem cell characteristics . More importantly, our fatemapping study shows for the first time that there is a marked decline in replenishment of Gpr17+ COPs from OPCs, with a subsequent loss of MOLs. Gpr17 can be activated by uracil nucleotides and cysteinyl leukotrienes, whose brain levels increase upon damage, although under excessive inflammatory conditions, such as those also found in the aged brain, Gpr17 can be pathologically activated by oxysterols and stromal cell-derived factor-1 (SDF-1) (Fumagalli et al., 2011;Parravicini et al., 2016). At early differentiation stages, Gpr17 delays OPCs differentiation, via activation of Gα i/o and inhibition of cAMP-PKA (Simon et al., 2016), whereas at later maturation stages, Gpr17 removal from the membrane via ligand-induced desensitisation by G-protein receptor kinase phosphorylation is necessary for terminal differentiation of COPs (Daniele et al., 2014;Fumagalli et al., 2011). By binding to Gpr17, inflammatory molecules could disrupt stage-dependent Gpr17 regulatory mechanisms, thus resulting in impaired COP terminal maturation and myelination. Since Gpr17 is expressed on the cell membrane, and thus amenable for pharmacological manipulation, we envisage that novel selective molecules directly acting at the Gpr17 receptor level could revert ageing associated effects on myelination (Parravicini et al., 2020). Notably, in line with the above findings and with Gpr17 function as a sensor for brain damage (Lecca et al., 2008), antagonism of Gpr17 has a rejuvenation effect in the ageing brain (Marschallinger et al., 2015). Our chromogenic immunohistochemical data and Gpr17 fate-mapping in Cre-Lox mice demonstrate major disruption of Gpr17 at both the mRNA and protein level in aged OPCs. In addition, we identified novel interactions in Gpr17 that are altered in ageing OPCs, with a prominent interaction with Gng10 (G-Protein Subunit Gamma 10), which links Gpr17 to both OPC proliferation and negative regulators of OPC differentiation, namely Pdgfra, Sox 4, Sox6 and Egfr (Baroti et al., 2016;Braccioli et al., 2018;Ivkovic et al., 2008). Our data identify a pivotal role for Gpr17 dysregulation in the ageing brain and the decline in OPC capacity to regenerate oligodendrocytes.

| Pharmacogenomicscreeningidentifies LY294002asatherapeutictargetforstimulating OPCsinthecontextofremyelinationinoldermice
We employed two separate pharmacogenomic approaches for determining: (a) the most optimal therapeutic agents for enhancing the densities of OPCs in the CC and their terminal differentiation into oligodendrocytes; (b) small molecules capable of reshaping aged transcriptional networks into their younger counterparts where their efficiency for myelin generation is more pronounced. Small molecules obtained in our analysis included those that target mTOR regulated cellular processes, including lipid metabolism, nucleotide synthesis and translation (Figlia et al., 2018), and, more specifically for this study, Gpr17 signalling and OPC maturation (Fumagalli et al., 2015;Ren et al., 2012). The PI3K/Akt/mTOR signalling inhibitor LY294002 was identified as the most potent small molecule for shifting the transcriptional hallmarks of aged OPCs into those characteristic of younger OPCs. LY294002 target genes in rejuvenating OPCs included ECM reorganisation, which we show above is dysregulated in aged OPCs. Importantly, we demonstrate that LY294002 promotes regeneration of OPCs and oligodendrocytes in vivo following demyelination induced by the toxin lysolecithin in 6-month-old mice, at which age the pace of remyelination is markedly impaired compared to younger adults (Crawford et al., 2016;Kazanis et al., 2017). These results in vivo validate the in silico pharmacogenomic data and demonstrate that small molecules identified using this approach have considerable potential in reversing the decline in OPC function in ageing and promoting remyelination and repair, likely via multiple effects that may also include Gpr17 regulation. In this study, we did not specifically assess the rejuvenating effects of ligands directly acting on Gpr17 (Parravicini et al., 2020), which will represent the focus of future studies.

| Conclusions
Our unbiased transcriptomic analysis identified oligodendroglial genes amongst the most altered in the aged mouse cerebrum, highlighting Gpr17 as a major factor in the disruption of the regenerative capacity of OPCs and decline in myelination. Unravelling the key transcriptional networks and signalling pathways that are central to age-related dysregulation of OPCs, enabled us to pharmacogenomically stimulate OPC rejuvenation in the context of remyelination.
Finally, it should be noted that changes in neuroglia appear to be a general feature of ageing, with evidence that astrocytes, microglia and OPCs undergo cellular atrophy, with a concurrent disruption of function in the course of normal ageing (Streit et al., 2004;Vanzulli et al., 2020;Verkhratsky et al., 2020). These studies provide a framework for future investigations in the field for targeting cellular mechanisms underlying the decline in glial plasticity and highlight the power of systems biology tools for counteracting the age-related decline in regenerative capacities and pathology.

| Animalsandtissue
All animal studies were performed in accordance with interna-  (Rivers et al., 2008) and Gpr17-iCreERT2xCAG-eGFP mice (Viganò et al., 2016) were, respectively, maintained and bred at the University of Portsmouth and University of Milan facilities; offspring were ear punched and genotyped using PCR, as previously reported (Rivers et al., 2008;Viganò et al., 2016), and mice of the correct phenotype and age (see below for the ages used) were injected intraperitoneally (i.p.,) twice a day for 5d with tamoxifen (0.1 ml of a 10 mg/ml solution, prepared in ethanol and corn oil), to induce Cre recombination and reporter expression, and brains were examined 10 days after the last injections (see below). The experiments were designed in compliance with the ARRIVE guidelines and no mice were excluded from analyses and experimental groups contained a spread of sexes. Control groups were included in all experiments, applying randomising procedures and double-blinded analysis when possible.

| Immunohistochemistry
For immunohistochemistry, mice were perfusion fixed intracardially under terminal anaesthesia with 4% paraformaldehyde (PFA). Brains were then dissected free and immersion fixed in 4% PFA overnight.
After fixation, tissues were washed 3 times in PBS and stored at 4°C in PBS containing 0.05% NaN3 (Sigma) until use. Coronal brain sections were cut using a vibratome (Leica) at a thickness of 60 μm and

| Imagingandanalysis
Images were captured using a Zeiss Axiovert LSM 710 VIS40S confocal microscope and maintaining the acquisition parameters constant to allow comparison between samples within the same experiment. Acquisition of images for cell counts was done with x20 objective. Cell counts were performed in a constant field of view (FOV) of 100 µm × 100 µm or 200 µm × 200 µm, depending on the area analysed, in projected flattened images from z-stacks formed by 10 or 15 z-single plain images with 1 µm interval between them; cell density was calculated as the number of cells divided by the area of the region analysed. All data were expressed as Mean ± SEM and tested for significance using unpaired t tests.

| RNA-seq
For gene profiling, cerebral hemispheres from 1-and 18-month-old C57/BL10 mice were removed (n = 3 mice from each age), maintaining strict RNAase-free and sterile conditions throughout. RNA was extracted and processed using an RNeasy Micro kit ( and the size profile was analysed on the 2200. Next, individual libraries were assessed, using the agilent 2200 bioanalyser for quality and quantity, and then normalised and pooled together accordingly.
The pooled library was denatured and loaded onto the HiSeq platform sequencer for paired-end sequencing.

| Meta-analysisofgenerateddatasets
Normalised datasets generated by RNA-seq were analysed with EdgeR using standard pipeline methods. Differential expression analysis datasets were further processed as described previously (Rivera & Butt, 2019), using ConsensusPathDB, STRING V10 (Herwig et al., 2016;Szklarczyk et al., 2015) and STITCH v5.05 (Kuhn et al., 2008). Normalised genes dysregulated in ageing were compared to cell-specific gene databases for OPC and MOL, using multiple published datasets (Zhang et al., 2014). The top 25 most significant genes associated with ageing in oligodendroglia are presented as a heatmap by ranking via FDR and relative fold change; the heatmap was constructed in R/Studio using the gg-plot2 package.

| Inductionofdemyelination
Mice aged 6 months were deeply anaesthetised under isofluorane and a cannula (Alzet, Brain infusion kit 3) was implanted at the coordinates Bregma −0.5 mm, lateral 1 mm, depth 2.5 mm, for intraventricular infusion of agents (Azim et al., 2017). Three days following implantation, mice were anaesthetised under isofluorane and 2% lysolecithin (LPC, Lα-lysophosphatidylcholine; Sigma-Aldrich) in a volume of 2 μl or sterile vehicle (saline/DMSO) was injected through the cannula to induce demyelination in the CC, as described previously (Azim & Butt, 2011). At 5 days post-injection (DPI) of LPC, the PI3 K/Akt inhibitor LY294002, or sterile vehicle (DMSO in saline) in controls, was delivered into the cerebrospinal fluid (CSF) for 4 days via the implanted cannula, using an osmotic minipump (10 μl/h, model 1003D; Alzet Osmotic Pumps); LY294002 was administered to provide a final concentration of 2 µM in the CSF, correcting for dilution in the ventricular volume, as described previously (Azim & Butt, 2011). To measure cell proliferation, mice were given an i.p.
injection of EdU (5-ethynyl-2′-deoxyuridine; Click-it EdU Alexa Fluor 488 imaging kit, Invitrogen) at 5 and 6 days DPI at 50 mg/kg. Brains were analysed at 10 DPI for gene expression and 14 DPI for cell analysis.

| qPCR
For real-time qPCR, total RNA was extracted as above in 4.4, from whole cerebrum to determine age-related changes in Gpr17 expression. To determine the effects of LY294002 on demyelination, as described above in 4.7, coronal brain sections of 500 µm thickness were cut using an adult mouse brain matrix and the CC was microdissected then flash-frozen (analyses were performed on n = 3/4 samples, each of 2 pooled CC), as described previously (Azim et al., 2017). Maintaining strict RNAase-free and sterile conditions throughout, total RNA extracted from these samples was reverse Amplification was performed using a Roche Lightcycler 96 according to the manufacturer's protocol. Data normalisation was performed using the housekeeping genes Gapdh and Rpl13a and expressed as relative gene expression using the 2ΔΔ-CT method.
Assays on all samples were performed in duplicate. Primer sequences are provided in Table S1.

| Statisticalanalysis
All statistical analyses were performed using GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA, USA) software. Data were expressed as mean ± standard error of the mean (SEM). The groups were compared using two-tailed Student's t test where appropriate and p < 0.05 was considered as statistically significant.

ACK N OWLED G EM ENTS
We would like to thank Dr Davide Marangon for perfusing mice (Milan).

CO N FLI C TO FI NTE R E S T S
Prof Arthur Butt and Dr Andrea Rivera are shareholders and cofounders of the company GliaGenesis. All authors declare no other conflicts.

DATAAVA I L A B I L I T YS TAT E M E N T
The transcriptomic datasets that support the findings of this study are available in the following link: https://uni-duess eldorf.sciebo.
de/s/72mMg xe40W 6iTFS and upon acceptance will be placed in the Github repository. All other data that support the findings of this study are available from the corresponding authors upon reasonable request.