Cell-Type-Specific Intralocus Interactions Reveal Oligodendrocyte Mechanisms in MS

Daniel C Factor; Anna M Barbeau; Kevin C Allan; Lucille R Hu; Mayur Madhavan; An T Hoang; Kathryn EA Hazel; Parker A Hall; Sagar Nisraiyya; Fadi J Najm; Tyler E Miller; Zachary S Nevin; Robert T Karl; Bruna R Lima; Yanwei Song; Alexandra G Sibert; Gursimran K Dhillon; Christina Volsko; Cynthia F Bartels; Drew J Adams; Ranjan Dutta; Michael D Gallagher; William Phu; Alexey Kozlenkov; Stella Dracheva; Peter C Scacheri; Paul J Tesar; Olivia Corradin

doi:10.1016/j.cell.2020.03.002

. Author manuscript; available in PMC: 2021 Apr 16.

Published in final edited form as: Cell. 2020 Apr 3;181(2):382–395.e21. doi: 10.1016/j.cell.2020.03.002

Cell-Type-Specific Intralocus Interactions Reveal Oligodendrocyte Mechanisms in MS

Daniel C Factor ^2,⁸, Anna M Barbeau ^1,⁸, Kevin C Allan ², Lucille R Hu ², Mayur Madhavan ², An T Hoang ¹, Kathryn EA Hazel ¹, Parker A Hall ¹, Sagar Nisraiyya ¹, Fadi J Najm ², Tyler E Miller ^2,^3,⁴, Zachary S Nevin ², Robert T Karl ², Bruna R Lima ¹, Yanwei Song ¹, Alexandra G Sibert ¹, Gursimran K Dhillon ², Christina Volsko ⁵, Cynthia F Bartels ², Drew J Adams ², Ranjan Dutta ⁵, Michael D Gallagher ¹, William Phu ¹, Alexey Kozlenkov ^6,⁷, Stella Dracheva ^6,⁷, Peter C Scacheri ^2,³, Paul J Tesar ², Olivia Corradin ^1,^9,^*

PMCID: PMC7426147 NIHMSID: NIHMS1581397 PMID: 32246942

SUMMARY

Multiple sclerosis (MS) is an autoimmune disease characterized by attack on oligodendrocytes within the central nervous system (CNS). Despite widespread use of immunomodulatory therapies, patients may still face progressive disability because of failure of myelin regeneration and loss of neurons, suggesting additional cellular pathologies. Here, we describe a general approach for identifying specific cell types in which a disease allele exerts a pathogenic effect. Applying this approach to MS risk loci, we pinpoint likely pathogenic cell types for 70%. In addition to T cell loci, we unexpectedly identified myeloid- and CNS-specific risk loci, including two sites that dysregulate transcriptional pause release in oligodendrocytes. Functional studies demonstrated inhibition of transcriptional elongation is a dominant pathway blocking oligodendrocyte maturation. Furthermore, pause release factors are frequently dysregulated in MS brain tissue. These data implicate cell-intrinsic aberrations outside of the immune system and suggest new avenues for therapeutic development.

Graphical Abstract

graphic file with name nihms-1581397-f0006.jpg

In Brief

Intralocus genetic risk factors reveal pathogenic cell types of individual disease-associated alleles and identify an oligodendrocyte-intrinsic dysfunction that confers risk to multiple sclerosis.

INTRODUCTION

Multiple sclerosis (MS) is a chronic disease involving central nervous system (CNS) infiltration of self-reactive lymphocytes, demyelination, and subsequent disability (Bennett and Stüve, 2009; Hafler et al., 2005; McFarland and Martin, 2007). More than 200 MS risk loci have been identified through genomewide association studies (GWASs) (Hafler et al., 2007; De Jager et al., 2009; Beecham et al., 2013; Patsopoulos et al., 2011; Sawcer et al., 2011). Historically, these loci have largely implicated immunological target genes such as major histocompatibility complex (MHC), interleukins (IL-2 and IL7R), and tumor necrosis factor receptor 1 (TNFR1) and pathways such as T cell differentiation and leukocyte activation (Baranzini and Oksenberg, 2017; Dendrou et al., 2015; Sawcer et al., 2011). 90% of MS risk loci map to non-coding regions of the genome and are strongly enriched in transcriptional enhancer elements active in T cells, further supporting a primary role for immune response in MS pathogenesis (Cusanovich et al., 2018; Farh et al., 2014; Hnisz et al., 2013; Maurano et al., 2012). Over the last decade, this T-cell-centric model has evolved to include additional cell types such as B cells, myeloid cells, and resident immune cells in the CNS (Filippi et al., 2018; Gandhi et al., 2010). Recently, the International Multiple Sclerosis Genetics Consortium completed a tour-de-force effort to expand the genetic map of MS, reporting genetic variants that account for ~48% of MS heritability. In addition to T and B cells, these risk loci implicate innate immune cells and microglia (Gosselin et al., 2017; International Multiple Sclerosis Genetics Consortium, 2019).

Oligodendrocytes also play a critical role in the progression of MS. In relapsing-remitting MS, patients experience acute periods of inflammation leading to demyelination and neurological symptoms, followed by periods of remission (Dendrou et al., 2015). During early stages of the disease, remyelination occurs to some degree, but it ultimately fails, leaving unmyelinated axons (Keirstead and Blakemore, 1999). Oligodendrocyte progenitor cells (OPCs) are essential for the generation of myelin sheaths. OPCs are present throughout the CNS, even in MS patients (Chang et al., 2000,2002), but fail to replace the significant numbers of lost oligodendrocytes. Current MS therapies are aimed at preventing further immune-mediated damage of myelin. While these drugs decrease relapse rates, they ultimately fail to slow long-term disease progression or avert neurodegeneration, though the long-term effects of early immunomodulatory intervention are not yet clear (Feinstein et al., 2015; Haghikia et al., 2013). Further insights into the genetic contributions to MS risk have the potential to reveal whether risk SNPs affect genes with functional commonalities beyond autoimmunity and augment immune-centric therapeutic paradigms. While efforts are underway to develop neuroprotective or remyelination therapies, the contribution of CNS-intrinsic defects to MS risk remains unclear.

We developed a strategy to expand understanding of GWAS results by identifying cell types impacted by the presence of a given disease allele. Like many complex traits, MS risk variants are profoundly enriched in enhancer clusters, genetic loci characterized by a high density of active regulatory elements that physically interact in the context of the 3-dimensional genome (Corradin et al., 2014; Hnisz et al., 2013). Functional studies in model organisms and saturation mutagenesis experiments with CRISPR/Cas9 have shown that these regions often contain multiple functional enhancers that collectively contribute to gene regulation (Canveret al., 2017; Corradin et al., 2014; Fortini et al., 2014; Glubb et al., 2015; Guo et al., 2015; He et al., 2015; Rajagopal et al., 2016). We previously investigated the combinatorial role of DNA variants within enhancer clusters in defining target gene expression and altering risk to disease. We identified “outside variants,” DNA variants that physically interact with the same gene target as GWAS SNPs but are in low linkage disequilibrium (LD) with the reported GWAS SNP. Despite the low LD of outside variants and GWAS SNPs, we identified outside variants that collude with risk SNPs to influence target gene expression and significantly impact disease risk. Additionally, including outside variants in estimates of heritability increased total heritability explained by GWAS loci 3- to 5-fold (Corradin et al., 2016).

Here, we identify outside variants that physically interact with the same target gene as MS GWAS SNPs and cooperatively contribute to MS risk. For each GWAS locus, outside variants provide additional information about the cis regulatory elements that are critical to disease pathogenesis. Our approach integrates these “intralocus” genetic risk factors with cell-type-specific chromatin activity to identify the pathogenic cell type of individual MS risk loci (Figure 1A). The advantage of this approach over conventional “enrichment-based” strategies is that it facilitates the identification of relevant cell types for GWAS SNPs on a locus-by-locus basis. Using this method, we identified an unexpected oligodendrocyte intrinsic contribution to MS risk via disruption of key genes involved in the regulation of RNA polymerase II (RNAPII) release. Our data revealed a functional role for transcriptional pause release regulation in the maturation of oligodendrocytes. Collectively, these results illustrate the power of our approach to reveal new insights into disease biology by identifying the pathogenic cell type of disease alleles.

Figure 1. — (A) Diagram of approach to identify pathogenic cell type of individual GWAS risk loci. Outside variants physically interact with the same target gene as a GWAS variant. A two-tiered stratification approach is used to determine the impact of outside variants on disease risk (left). Genetic risk barcode is composed of outside-variant sites that significantly alter clinical risk compared to sites where outside variants have no contribution to risk (center). Cell types with H3K27ac molecular activity that is concordant with the genetic risk barcode are identified as likely pathogenic (right).

(B) Example GWAS SNP rs13333054, highlighted in gray, physically interacts with *IRF8* and *EMC8* in activated CD4⁺ T cells (purple) and monocytes (blue). Arcs represent physical interactions identified via promoter capture Hi-C. Highlighted in red is an exemplar outside-variant site that significantly alters risk and is enriched for H3K4me1/H3K27ac and Hi-C activity in monocytes, but not in T cells.

(C) GWAS SNPs overlap with H3K27ac peaks and are row ordered based on lineage-specific clusters (far right). Cell-type columns are highlighted if GWAS and/or LD SNPs (r² > 0.8) overlap H3K27ac peaks identified in each cell type (left, top). MS GWAS results for genome-wide chromatin activity enrichment via Variant Set Enrichment (VSE) and LD score regression (LDSR) for MS (left, bottom). Pathogenic cell types identified by outside-variant approach are shown. For each GWAS locus (row), the identified cell types are highlighted (right). Clusters of GWAS loci, as defined by the outside-variant predictions, were identified via similarity of cell type and lineage specificity (far right).

(D) Comparison of lineage-specific clusters identified by an outside-variant approach to the heritability of 28 other disorders via LDSR. Heritability enrichment of six other autoimmune disorders within each lineage-specific cluster is shown (left). The top six traits with the strongest heritability enrichment in loci identified to act in the CNS for MS are shown (right). The multi-test-corrected significance threshold is denoted by a dashed line.

See also Figures S1, S2, and S3 and Tables S1, S2, and S3.

RESULTS

Cell-Type Specificity of Intralocus Regulatory Elements Enables Prediction of Pathogenic Cell Type of MS Risk Loci

We developed an approach to identify the pathogenic cell type of individual disease-associated alleles by integrating intralocus genetic risk factors with cell-type-specific chromatin activity. Candidate outside variants, SNPs that physically interact with the same target gene as MS GWAS SNPs, were identified using Hi-C data from 15 immune cell types as well as neural precursor cells (NPCs) (Figure S1). We curated a panel of publicly available ChIP-seq experiments from 20 immune and CNS cell types and filtered candidate outside variants to those found in putative regulatory elements marked by H3K27ac enrichment (Figure S2; Table S1). In order to evaluate the impact of candidate outside variants on MS disease risk, we utilized a two-tier stratification approach that we previously described (Corradin et al., 2016) (Figure 1A, left). Here, individuals are first stratified by genotype of GWAS SNP and then again by genotype of outside variant, resulting in up to nine possible genotype combinations. Each genotype combination is evaluated separately to identify outside variants that significantly contribute to disease risk beyond the effect of the GWAS SNP (STAR Methods; Figure S1). We identified outside variants that significantly altered MS risk for 126 out of 163 MS GWAS loci.

We reasoned that outside variants provide novel information about the intralocus genomic regions that contribute to disease risk and would enable us to distinguish local regulatory regions that alter risk from those that do not (Figure 1A, center). We hypothesized that comparing these genetic risk “barcodes” to cell-type-specific molecular activity, such as active chromatin features (H3K27ac), could facilitate the identification of the cell type(s) in which risk alleles exert their pathogenic effect. For this approach, cell types for which active molecular activity coincided with regions that contributed to disease risk would be predicted as likely pathogenic (Figure 1A, purple). Conversely, cell types for which molecular activity did not coincide with genetic risk (Figure 1A, blue/green) would be less likely to be defined as pathogenic for a given disease locus.

An example is the IRF8 locus shown in Figure 1B. Here, an MS risk SNP (highlighted in gray) lies within putative regulatory elements identified by histone modifications H3K4me1 and H3K27ac. Furthermore, the enhancer harboring the risk SNP was determined through promoter capture Hi-C to physically engage promoters in nine immune cell types, including activated CD4⁺ T cells (purple) and monocytes (blue) (Javierre et al., 2016). We identified outside variants that significantly altered risk to MS, beyond the effect of the reported GWAS SNP (Figure 1B, top). An example outside-variant site is highlighted in red. In contrast to the GWAS site, the molecular activity at this outside-variant site is highly cell-type specific, with H3K4me1/H3K27ac and Hi-C activity only identified in monocytes. Outside variant sites that influence disease risk and are involved in this type of cell-type-specific intralocus interaction provide us with an opportunity to distinguish likely pathogenic cell types (STAR Methods).

We applied this approach to the 126 MS risk loci for which we identified outside variants and predicted pathogenic cell types for 104. This included loci predicted to act outside of T and B cells and, more surprisingly, outside of the immune system altogether. We grouped the GWAS loci based on the pathogenic cell type predicted by the outside-variant approach; e.g., loci predicted to act primarily in the T cells were grouped together, and those predicted to act in monocytes and macrophages were grouped as myeloid predictions (Figure 1C, right; Table S2). We compared these results to the standard approach of overlapping the GWAS SNP (and SNPs in LD) with active regulatory elements identified by chromatin immunoprecipitation sequencing (ChIP-seq). While MS GWAS SNPs are significantly enriched in T cell active regulatory elements (Figure 1C, bottom left), the majority of GWAS SNPs also overlap with H3K27ac across a wide variety of cell types, including B cells, monocytes, and macrophages (Figure 1C, top left). In contrast, for the majority of MS risk loci, the outside-variant approach was able to narrow down specific cell types likely to be pathogenically impacted by the risk SNPs (Figure 1C, right). This included not only 35 loci predicted to act primarily in T cells but also 6 other loci predicted to impact B cells and 18 loci in myeloid cells. Among the risk loci predicted to affect myeloid cells was the IRF8 locus highlighted in Figure 1B. This prediction is well supported by a previous study that demonstrated that Irf8 expression in monocytes, but not in T cells, is required for the progression of MS in the mouse experimental allergic encephalomyelitis (EAE) MS model (Yoshida et al., 2014). In addition to loci that act in the immune system, we also identified six MS loci to be pathogenic specifically in the CNS. Thus, incorporating outside variants and intralocus enhancer-gene interactions enables us to identify loci that may act outside of the proposed disease causal cell types, which was not possible with previous enrichment-based strategies.

To validate these results, we utilized two diverse approaches: functional perturbation of outside variants and evaluation of genetic similarity of our outside-variant prediction clusters with other human traits. Our approach is based on the premise that outside variants found in H3K27ac-enriched sites will functionally contribute to gene regulation in the identified cell type. We tested this premise by utilizing CRISPR/Cas9 to generate indels at outside-variant sites and assessing the impact of mutations on target gene expression in a T cell model (Jurkat cells) and a B cell model (GM12878). We selected a locus that was identified to act specifically in T cells (rs7191700). While indels at the transcription start site (TSS) decreased expression of target gene SOCS1 in both cell types, mutations generated at outside-variant sites decreased expression of the target gene in the T cell model, while mutations in the same loci induced little to no effect on expression in the B cell model (Figure S3; Table S3). This supports not only that outside-variant sites functionally contribute to gene regulation but also that our approach can distinguish cell-type-specific effects.

We next sought to globally assess our predictions by evaluating which MS genetic risk factors contribute to the heritability of other diseases and traits (Bulik-Sullivan et al., 2015; Finucane et al., 2015. Autoimmune disorders, including MS, have a strong shared genetic architecture (Cotsapas et al., 2011; Beecham et al., 2013). Risk loci identified by GWASs often confer risk to multiple autoimmune traits. Thus, we hypothesized that MS risk loci that act in the T cells would be most likely to explain heritability for other autoimmune disorders, while any MS risk loci that act outside of the immune system would be unlikely to coincide with genetic risk to autoimmune disorders. We found that the loci predicted to act in the T cells explained the largest portion of genetic similarity with other autoimmune disorders, while CNS predicted loci explained minimal heritability of other autoimmune disorders. This supports our predictions, as risk factors that act in the CNS in MS are the most likely to distinguish MS from other autoimmune disorders (Figure 1D, left).

We applied this approach to a total of 28 disorders (Table S1) and found that the MS risk loci predicted to act in the CNS had the strongest similarity with traits with known CNS function, including depression, autism, and body mass index (Locke et al., 2015) (Figure 1D, right). Importantly, none of these traits share a strong genetic correlation with MS. The genetic overlap was specific to the MS risk loci we predicted to act in the CNS. Together, these results demonstrate that outside variants can be utilized to define likely pathogenic cell types of individual risk loci and furthermore identify groups of risk loci that act within the same cell type or lineage.

Oligodendrocyte-Specific Regulation of Transcriptional Pause Release Factors Are Associated with MS Risk

Despite failure of remyelination being a dominant cellular pathology in MS patients, an oligodendrocyte-intrinsic contribution to MS risk or severity has not yet been recognized. Using our outside-variant approach, we identified three loci to be pathogenic specifically in the oligodendrocyte lineage. This prediction was consistent across three biological replicates of ChIP-seq of oligodendrocytes isolated from independent human donors (Kozlenkov et al., 2018) (Figures 2A–2D). We identified putative target genes using publicly available Hi-C data of upstream neural progenitors and prefrontal cortex (Schmitt et al., 2016). For two out of three loci, we found that GWAS SNPs and outside variants physically interacted with target genes that play a key role in the regulation of transcriptional elongation via the release of promoter-proximal RNAPII. This included a BET-bromodomain-containing protein BRD3 (Figures 2C and 2D), which facilitates transcriptional elongation (Dai et al., 2019; LeRoy et al., 2008), and HEXIM1/2 (Figures 2A and 2B), which inhibits transcriptional elongation by sequestering P-TEFb in an inactive complex (Jonkers and Lis, 2015; Young et al., 2007). We found that several outside-variant sites that are active enhancers in the oligodendrocytes to also be active in neuronal precursor cells (NPCs) and thus leveraged NPCs to assess the functional impact of these regions on gene regulation. Mutations at outside-variant sites induced by CRISPR/Cas9 led to a significant reduction in HEXIM1 (Figure 2E) and BRD3 (Figure 2F) mRNA levels at each respective locus, demonstrating the functional role these sites play in regulating the transcription of pause release factors.

Figure 2. — (A) *HEXIM1/2* locus outside variant results (top, Manhattan plot). H3K27ac enrichment across multiple cell types, including three biological replicates of oligodendrocytes (bottom). Highlighted in red are exemplar outside-variant sites that significantly alter risk and are enriched for H3K27ac in oligodendrocytes.

(B) Aggregate plot of oligodendrocyte H3k27ac enrichment at intralocus outside-variant sites at the *HEXIM1/2* locus. H3k27ac signal aggregate across sites that alter risk (solid line) versus sites that do not alter risk are shown for three ChIP biological replicates. *kruskal-Wallis one-sided p < 2E-6 for each oligodendrocyte ChIP.

(C) Same as in (A), but for the *BRD3* locus.

(D) Same as in (B), but for the *BRD3* locus.

(E) *HEXIM1* gene expression measured by qRT-PCR in an NPC model following transfection of Cas9 and sgRNA targeting outside variant sites (sg1-sg4) or *HEXIM1* TSS. Log₂ fold change for two biological replicates shown for each guide compared to cells transfected with Cas9 without a gRNA.

(F) Same as in (E), but for the BRD3 locus.

(G) MS heritability enrichment estimated for transcriptional pausing gene sets derived from Reactome (n = 200) and Gene Ontology (n = 100) using LDSR. The multi-test-corrected significance threshold is denoted by a dashed line.

See also Figure S3 and Tables S1, S3, and S4.

These observations led to the hypothesis that dysregulation of transcriptional elongation in the oligodendrocyte lineage contributes to MS pathogenesis. Transcriptional pausing is a critical mechanism by which the cell regulates gene expression. A growing number of genes have been linked to the regulation of promoter-proximal RNAPII pausing (Jonkers and Lis, 2015). We aimed to assess the genetic contribution of this pathway to MS, independent of the outside-variant approach. We utilized LD score regression (LDSR) to assess MS heritability enrichment for two curated gene sets derived from Reactome (Fabregat et al., 2018) and Gene Ontology (GO) (see STAR Methods) (Finucane et al., 2015). Both transcriptional pausing gene sets were significantly enriched for MS heritability, supporting the hypothesis that genetic variation in this pathway contributes to MS (Figure 2G; Table S4).

Inhibition of Transcriptional Pause Release Blocks Oligodendrocyte Maturation

Together, these results provide genetic support for dysregulation of pause release genes in the oligodendrocyte lineage as a basis for MS risk. We sought to utilize an unbiased approach to identify fundamental pathways that are critical to the maturation of oligodendrocytes and thus the generation of new myelin. Performing an unbiased screen enabled us to identify not only whether transcriptional pause release plays a role in oligodendrocyte maturation but also how this pathway compares to other pathways that are critical to the generation of new myelin.

We adapted a high-content image-based phenotypic screening platform with a proven record identifying modulators of oligodendrocyte maturation (Najm et al., 2011, 2015). A library of 3,141 compounds with well-annotated targets was screened to identify compounds that block or delay the generation of mouse oligodendrocytes from OPCs. Compounds that negatively impacted cell viability or oligodendrocyte lineage marker OLIG2 were excluded. Known pro-myelinating compounds benztropine, clemastine, ketoconazole, and miconazole increased the fraction of cells positive for the oligodendrocyte marker myelin basic protein (MBP), validating the overall performance of the screen (Deshmukh et al., 2013; Mei et al., 2014; Najm et al., 2015) (Figures 3A and S4A–S4C). Consistent with our outside-variant genetic analysis, we found four4 dual-bromodomain BET inhibitors, which potently inhibit transcriptional elongation, in the top 2% of compounds that inhibited oligodendrocyte maturation. In total, four out of five BET bromodomain inhibitors included in the screen were classified as hits at the single dose tested, making it the predominant pathway identified in our screen to inhibit oligodendrocyte maturation (Figure 3A).

Figure 3. — (A) A library of 3,141 bioactive compounds was screened at 2 μM to identify inhibitors of development of mouse MBP-positive oligodendrocytes. Data represent log₂ fold change from standard conditions promoting development (T3; dotted line). Baseline maintenance conditions of OPCs are indicated by the fibroblast growth factor (FGF) dotted line. Select compounds known to induce development are highlighted in yellow. The 2% of compounds that best prevented oligodendrocyte development are black, while dual BET bromodomain inhibitors are highlighted in blue.

(B) Quantification of MBP with varying culture conditions. The inhibitory action of dual BET bromodomain inhibitor JQ1 is limited to the racemic form and active (S)- stereoisomer, indicating the effect is likely on-target. Each point represents the average of five fields from a single well. ns, p > 0.05; ****p < 0.0001 by Dunnett’s multiple comparisons test.

(C) Dose response of the effect of (S)-JQ1 on formation of MBP-positive oligodendrocytes. JQ1 inhibits oligodendrocyte maturation with an IC₅₀ of 105 nM (dotted line).

(D) Representative immunostaining of mouse OPCs grown in oligodendrocyte-promoting conditions and either inactive (R)-JQ1 or active (S)-JQ1 for 3 days or for 6 days with (S)-JQ1 washout at 3 days. OLIG2 (red) marks cells of the oligodendrocyte lineage, while MBP (green) indicates mature oligodendrocytes. Scale bars indicate 50 μm.

(E) Representative immunostaining of mouse OPCs co-cultured with DRG neurons for 7 days in the presence of either inactive (R)-JQ1 or active (S)-JQ1 or for 10 days with (S)-JQ1 washout at 3 days. Neurofilament (NF) is shown in red and MBP in green. Scale bars indicate 100 μm.

See also Figures S4 and S5.

To further evaluate the role of BET bromodomain inhibition on myelination, we utilized the canonical small-molecule inhibitor JQ1 (Filippakopoulos et al., 2010). Racemic JQ1 and the active (S)- stereoisomer both dramatically inhibited cell maturation in a dose-dependent manner, while expression of OLIG2 was maintained (Figures 3B–3D and S4D). In contrast, vehicle and the inactive (R)- stereoisomer did not inhibit the maturation of oligodendrocytes, confirming the on-target effect of JQ1 (Figures 3B and 3D). We next co-cultured OPCs with rat dorsal root ganglia (DRG)-derived neurons in order to test the impact on myelination in more physiologically relevant conditions. We found that BET bromodomain inhibition prevented OPCs from developing and associating with neural axons (Figures 3E and S4E). We found the effect of JQ1 to be reversible in each of these experiments, as removing JQ1 enabled cells to mature into myelinating oligodendrocytes (Figures 3D, 3E, and S4E). This suggests that JQ1 did not divert the cells to an alternative cell state; rather, JQ1 temporarily blocked maturation. Finally, since our primary OPC screening platform utilized mouse cells, we validated that JQ1 also inhibits the generation of human oligodendrocytes using OPCs derived from human induced pluripotent stem cells (iPSCs) (Figures S4F and S4G). Together, these findings demonstrate that BET bromodomain inhibitors have a substantial functional impact on the generation of new myelinating oligodendrocytes.

To determine whether JQ1’s effect on oligodendrocyte maturation is specifically attributable to its impact on transcriptional elongation, we compared gene expression in OPCs in maintenance conditions to that of cells cultured for 3 days in pro-maturation conditions with or without JQ1 (Figure S5A). We found that genes downregulated as OPCs differentiate to oligodendrocytes were also concordantly downregulated in the presence of JQ1. Likewise, genes that have consistent expression in OPCs and oligodendrocytes were not significantly altered by JQ1 treatment (Figures S5B–S5E). In contrast, genes that are normally induced during oligodendrocyte development, such as the proteolipid protein 1 (P/p7), failed to be activated in the presence of JQ1 (Figures S5F and S5G). Thus, BET bromodomain inhibition hinders oligodendrocyte maturation by preventing transcriptional activation of oligodendrocyte-specific genes. We evaluated RNAPII occupancy at genes inhibited by JQ1 using RNAPII ChIP-seq. In the presence of JQ1, we found a decrease in RNAPII occupancy at the gene bodies of oligodendrocyte-specific genes, implicating inhibited transcriptional elongation as the mechanism by which JQ1 prevents activation of the oligodendrocyte gene program (Figures S5H–S5J).

To more broadly evaluate whether perturbation of the transcriptional elongation blocks oligodendrocyte maturation, we tested whether inhibition of other proteins in the transcriptional elongation pathway resulted in similar phenotypes. Inhibition of either the P-TEFb component CDK9 or the histone methyltransferase DOT1L also blocked oligodendrocyte maturation, supporting the conclusion that transcriptional elongation is a dominant pathway blocking oligodendrocyte generation and myelination from OPCs (Figures S5K and S5L).

Dysregulation of Hexim1 Alters Oligodendrocyte Maturation

Our outside-variant approach led to the hypothesis that dysregulation of transcriptional elongation in oligodendrocytes contributes to the pathogenesis of MS. We found that chemical inhibition of this pathway impacted the maturation of oligodendrocytes. We next sought to assess whether dysregulation of Hexim1, acritical gene target in one of our predicted loci with a known role for inhibiting transcriptional elongation (Jonkers and Lis, 2015; Young et al., 2007), is sufficient to alter oligodendrocyte maturation. We utilized CRISPR interference (CRISPRi) and CRISPR activation (CRISPRa) to alter Hexim1 transcript levels. We designed gRNAs to target the Hexim1 promoter with catalytically dead Cas9 fused to either Kruppel-associated box (dCas9-KRAB) or VP64 (dCas9-VP64) in mouse OPCs to facilitate Hexim1 knockdown and Hexim1 activation, respectively.

OPCs transduced with CRISPRi and CRISPRa constructs were cultured for 3 days in pro-maturation conditions using the same approach as our chemical screen. CRISPRa targeted to the Hexim1 promoter led to a 2-fold increase in Hexim1 gene expression relative to the non-targeting control (NTC) gRNA, whereas CRISPRi led to a 4-fold decrease in Hexim1 expression (Figures 4A and 4D). We next assessed whether the alteration in expression was sufficient to disrupt oligodendrocyte maturation. We found that CRISPR-mediated activation led to a 50% reduction in MBP-positive cells (2B,C), whereas, CRISPR-mediated knockdown of Hexim1 resulted in a 8-fold increase in the portion of MBP-positive cells (Figures 4E and 4F). As Hexim1 is an inhibitor of transcriptional elongation, this is directionally consistent with the results of our drug screen that indicate inhibition of transcriptional elongation reduces the generation of mature oligodendrocyte from OPCs. These results suggest that dysregulation of Hexim1 alone is sufficient to alter oligodendrocyte maturation from OPCs. Furthermore, this supports our hypothesis that dysregulation of transcriptional elongation in the oligodendrocytes contributes to MS risk by inhibiting the maturation of oligodendrocytes.

Figure 4. — (A) qRT-PCR showing CRISPRa-mediated increase in *Hexim1* gene expression relative to a non-targeting control (NTC) and normalized to endogenous control *Rpl13a* (n = 4 technical replicates, mean ± SEM).

(B) Fold change of the percentage of oligodendrocytes (MBP/DAPI) after transduction of CRISPRa with guide targeting *Hexim1* relative to percentage of oligodendrocytes (MBP/DAPI) after transduction of CRISPRa-NTC (n = 14 independent wells; mean ± SEM; *p < 0.002 by Wilcoxon test).

(C) Representative images of oligodendrocyte formation from CRISPRa-NTC, CRISPRa-Hexiim1 OPCs after maturation for 3 days. Myelinating oligodendrocytes are immunostained with anti-MBP (green) and cell nuclei are stained with DAPI (blue). Scale bars indicate 100 μm.

(D) qRT-PCR showing a CRISPRi-mediated decrease in *Hexim1* gene expression relative to an NTC and normalized to endogenous control *Rpl13a* (n = 4 technical replicates; mean ± SEM).

(E) Fold change of the percentage of oligodendrocytes (MBP/DAPI) after transduction of CRISPRi with guide targeting *Hexim1* relative to percentage of oligodendrocytes (MBP/DAPI) after transduction of CRISPRi-NTC (n = 23 independent wells; mean ± SEM; **p < 0.0001 by Wilcoxon test). Same as in (C), but for CRISPRi-NTC and CRISPRi-Hexim1 representative images

See also Table S3.

Transcriptional Elongation Factors Are Dysregulated in MS Patient Brains

To further evaluate the role of transcriptional pausing in MS, we evaluated the expression of these genes in previously published MS patient tissue microarray data (Fluynh et al., 2014). Strikingly, BRD3 expression was significantly downregulated in white matter (WM) lesions of MS patients and HEXIM1 and HEXIM2 expression was significantly increased in MS patient WM lesions compared to controls (Figure 5A). This is directionally consistent with increased transcriptional pausing as BRD3 is an activator of transcriptional elongation and FIEXIM proteins inhibit elongation.

Figure 5. — (A) Microarray expression (RMA normalized) of elongation factors in patient white matter (WM). *HEXIM1, HEXIM2*, and *BRD3* are significantly dysregulated in WM lesions compared to control tissue by Dunnett’s multiple comparisons test.

(B) Representative immunostaining of PLP1 and HEXIM1 in control and MS patient tissue sections. Lesions are identified by PLP1-low regions in MS images. Scale bars indicate 50 μm.

(C) Quantification of HEXIM1/PLP1 double-positive cells in control tissue as compared to primary- and secondary-progressive MS. Each plotted point represents quantification of a single 0.07-mm² field. Co-localization significantly increases in both disease states by Dunnett’s multiple comparisons test, ns, p > 0.05; *p < 0.05; **p < 0.01; ****p < 0.0001.

See also Table S4.

Given the potential confounders of gene expression analysis in bulk MS lesion tissue such as immune infiltration and oligodendrocyte death, we next examined patient tissue sections immunostained for FIEXIM1 to determine whether dysregulation occurred in the oligodendrocyte lineage. Comparison of sections from patient lesions showed an increase of FIEXIM1 positive cells as compared to healthy controls (Figure 5B; Table S5). We utilized PLP1, a marker of early or mature oligodendrocyte identity, to evaluate FIEXIM1 expression specifically in oligodendrocyte lineage cells. We quantified the number of PLP1-positive oligodendrocyte lineage cells that also expressed FIEXIM1 in patient lesions compared to controls. Patients with primary progressive MS displayed a 6-fold increase in the number of cells co-stained for FIEXIM1 and PLP1 as compared to normal controls. Similarly, secondary progressive MS patient lesions showed a 10-fold increase in HEXIM1/PLP1 double-positive cells (Figure 5C). These data provide further evidence for cell-intrinsic defects in the transcriptional elongation pathway of the MS patient oligodendrocyte lineage.

DISCUSSION

Here, we describe an approach to identify the pathogenic cell type of individual risk loci. This approach differs from previous methods that utilize results from GWASs to identify disease critical cell types, as it evaluates each locus independently. This enables identification of clusters of risk loci that contribute to disease via dysregulation of the same cell type and thus enables us to identify disease critical pathways that may be obscured when all risk loci are evaluated collectively. Here, we demonstrate the power of this approach by identifying risk loci that act via dysregulation of expression in the oligodendrocytes within a shared pathway.

We functionally validated these computational findings with an unbiased approach to identify the pathways critical to the maturation of oligodendrocytes. We found that disruption of transcriptional pause release is a dominant pathway critical for oligodendrocyte maturation and thereby the generation of new myelin. In patient brain tissue samples, we also found that these gene targets were frequently dysregulated in the oligodendrocyte lineage. Importantly, our findings do not contradict the autoimmune T-cell-centric model of MS pathogenesis. Rather, we implicate an additional oligodendrocyte-intrinsic aspect of MS risk. These results suggest that some individuals may be predisposed toward increased transcriptional pausing and therefore be less efficient at generating new myelinating oligodendrocytes from OPCs. Whether the effect of these variants on myelin is more pronounced in the context of MS remains unknown.

While many genetic risk loci have implicated lymphocyte dysregulation, relatively few have implicated CNS cells, including astrocytes (Ponath et al., 2018) and microglia (Gosselin et al., 2017; International Multiple Sclerosis Genetics Consortium, 2019). Despite the expected role of myelin and oligodendrocytes in the etiology of MS, to our knowledge, this study presents the first example of oligodendrocyte-intrinsic dysfunction contributing to MS genetic risk. We also predicted three MS risk loci to contribute to MS risk via dysregulation of gene expression in neurons. Among the potential gene targets nominated by physical interaction with these GWAS loci in brain Hi-C data are BRINP1 and AVIL (Jung et al., 2019). Interestingly, both of these genes have been implicated in the positive regulation of neuronal differentiation (GO Biological Processes, 2018). Future studies are required to evaluate the critical gene targets of these loci and their potential contribution to neuronal dysfunction in MS pathogenesis. Predicting the pathogenic cell type of these loci enables the generation of specific hypotheses that can guide future functional studies that reveal further insights into the pathogenesis of MS.

We predicted 98 MS risk loci to act via dysregulation of immune cells and 6 loci to contribute to MS through CNS-intrinsic functions. It is however important to note that our predictions may be underestimating the role of CNS cell types in MS risk. Our approach is reliant upon high quality epigenomic profiles,which can be challenging to derive for neuronal and glial cell types. As our understanding of the epigenetic landscape of the human brain expands (Jung et al., 2019), we will be able to assess the impact of MS risk loci on the function of cell types such as microglia and astrocytes. Disease- or context-specific ChIP-seq profiling could also aid identification of regulatory elements that contribute to disease risk specifically in the environmental context of the trait. Additionally, here we evaluated genetic variants that contribute to risk of developing MS. Genetic variants that alter oligodendrocytes and other CNS cell types may be more likely to contribute to other MS characteristics such as severity, relapse rate, or rate of progression. Further studies that characterize the genetic risk associated with these phenotypes could reveal additional roles of oligodendrocytes in MS pathogenesis.

Enrichment-based approaches nominate disease relevant cell types by comparing epigenomic profiles to genetic risk to identify cells types that are globally enriched for genetic contribution to disease. Multiple studies have demonstrated that regulatory elements active in T cells are enriched for MS genetic risk (Cusanovich et al., 2018; Farh et al., 2014; Flnisz et al., 2013; Maurano et al., 2012). Previous work by Farh et al. (2014) identified likely causal SNPs for autoimmune disease risk loci and compared these loci to ChIP-seq from a comprehensive panel of immune cell types. They found the likely causal SNPs for MS to be enriched for active regulatory elements found in multiple immune cells, with strongest enrichment among stimulated T helper cell populations and B cells. These enrichment-based strategies are powerful approaches that can reveal primary disease causal cell types essential to disease pathogenesis. The outside-variant approach complements these methods by providing the opportunity to identify risk loci that act outside of the primary disease causal cell type. Despite several limiting factors, such as the degree of regulatory element cell-type specificity, the accuracy by which ChIP-seq can distinguish active and inactive regulatory elements, and the extent of LD within each locus, we identified pathogenic cell types for 104 MS GWAS loci. Together, these results suggest that applying this approach to multiple diseases and cell types could facilitate rapid delineation of the pathogenic consequence of disease-associated variants.

STAR★METHODS

LEAD CONTACT AND MATERIALS AVAILABILITY

Further information and requests for resources and reagents can be directed to the Lead Contact, Olivia Corradin (corradin@wi.mit.edu). The reagents generated in this study are available without restriction.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines

Jurkat cells

Human, male Jurkat cells (clone E6-1; ATCC, Manassas, VA) were maintained in Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (FBS; ATCC) and penicillin (100 U/ml)-streptomycin (100 μg/ml; Thermo Fisher Scientific) at 1x10⁵ – 1x10⁶ cells/ml and maintained at 37°C with 5% CO₂.

GM12878 cells

Human, female GM12878 cells (Coriell Institute for Medical Research, Camden, NJ) were maintained in RPM11640,15% FBS, 2mM L-glutamine (GE Healthcare Life Sciences, Pittsburg, PA) and penicillin (100 U/ml)-streptomycin (100 μg/ml) at 2x10⁵ – 1x10⁶ cells/ml and maintained at 37°C with 5% CO₂.

Primary cell culture

Generation of neural progenitor cells

Neuronal differentiation of human, male and female embryonic stem cells (ESC) and induced pluripotent stem cells (iPSC) was performed similarly to previous work (Muffat et al., 2016). WIBR3 ESCs (Lengner et al., 2010), and iPSCs generated from fibroblasts (human, 88 year old female and 29 year old male; Coriell Institute for Medical Research) were dissociated from mouse embryonic fibroblasts (isolated from E13.5 embryos of male and female CD-1 mice (in accordance with protocols approved by Massachusetts Institute of Technology’s Committee on Animal Care); Charles River Laboratories, Wilmington, MA) with 1.5 mg/mL collagenase (Thermo Fisher Scientific). Cells were plated in Matrigel-coated 6 well plates (Corning Life Sciences, Oneonta, NY) at a density of ≥ 5x10⁶ cells per well in neuroglial differentiation media (NGD; (Muffat et al., 2016)) containing 10 ng/mL human recombinant FGF-Basic (AA 1-155) (Thermo Fisher Scientific), an additional 10 mM of human insulin (MilliporeSigma, Burlington, MA), 10 μM Rho-associated protein kinase (ROCK) inhibitor Y-27632 (MilliporeSigma), and 2.5 μM dorsomorphin (Thermo Fisher Scientific). Cells were rinsed with phosphate buffered saline (PBS) and received fresh NGD media containing FGF, insulin and dorsomorphin every day for ~8-12 days, at which point neural rosettes were visible throughout the well. Dorsomorphin was then removed, and cells were passaged with StemPro Accutase Cell Dissociation Reagent (Thermo Fisher Scientific) using a 1:1 split ratio, and then expanded for two passages plating ≥ 5x10⁶ cells per well. Cells were frozen at passage 3 at 10⁷ cells per cryovial in NGD media containing FGF, insulin and ROCK inhibitor, and each vial was thawed into one Matrigel-coated well in the same media for all experiments.

Neural progenitor cells (NPCs) were maintained in Neurobasal Medium (Thermo Fisher Scientific) supplemented with 0.5X Gem21 Neuroplex without Vitamin A (Gemini Bio-Products, West Sacramento, CA), 0.2% (w/v) AlbuMAX I lipid-rich bovine serum albumin (Thermo Fisher Scientific), 1X N2 NeuroPlex serum-free supplement (Gemini Bio-Products, #400–163), 0.05 M NaCl (Thermo Fisher Scientific), 1X sodium pyruvate (Thermo Fisher Scientific), 2 mM GlutaMax (Thermo Fisher Scientific), 3.5 ng/ml biotin (AnaSpec, Fremont, CA) dissolved in 1 M NaOH (MilliporeSigma), 10 μM ascorbic acid (MilliporeSigma), 0.017% DL-lacticacid syrup (MilliporeSigma), 10 ng/ml human recombinant FGF-Basic (AA 1-155), and 0.02 mg/ml human insulin at 1x10⁶-5x10⁶ cells/ml on Matrigel-coated plates and medium was replaced each day. For passaging, cells were dissociated with accutase and plated in the above media with 10 mM ROCK inhibitor Y27632. NPCs were supplemented with penicillin (100 U/ml)-streptomycin (100 μg/ml) and maintained at 37°C with 5% CO₂.

Mouse OPC culture

OPCs were cultured as described previously, with the exception that SHH was not used for maintenance or differentiation (Najm et al., 2011). Mice were handled under protocols approved by Case Western Reserve University School of Medicine’s Institutional Animal Care and Use Committee (IACUC). First, epiblast stem cells (EpiSCs) were isolated from 129S/SvEv male embryos (E3.5; The Jackson Laboratory, Bar Harbor, ME). For neural rosette formation, cells were plated in Knockout Dulbecco’s Modified Eagle Medium (DMEM; Thermo Fisher Scientific) supplemented with 20% Knockout Serum Replacement (Thermo Fisher Scientific), 2 mM GlutaMax, 1X nonessential amino acids (Thermo Fisher Scientific), 0.1 mM 2-mercaptoethanol (MilliporeSigma) (EpiSC basal medium) plus 100 ng/ml Noggin (R&D Systems), 20 mM SB431542 (MilliporeSigma) and 2 mM dorsomorphin (MilliporeSigma). After 24 hours, cells were fed with 1 part EpiSC basal medium and 1 part neural basal medium (DMEM/F12 (Thermo Fisher Scientific), 1Xhigh insulin N-2 supplement (R&D Systems), 1X B-27 without vitamin A (Thermo Fisher Scientific), and 2 mM GlutaMax) supplemented with 100 ng/ml Noggin, 20 μM SB431542, and 2 μM dorsomorphin. After 24 hours, cells were fed with neural basal medium plus 100 ng/ml Noggin. After an additional 24 hours, 10 mM all-trans retinoic acid (ATRA; MilliporeSigma) was added. For differentiation into OPCs, neural rosettes were passaged into Nunclon-Δ plates coated with poly-L-ornithine (MilliporeSigma) and laminin (MilliporeSigma). Cells were maintained in neural basal medium plus 20 ng/ml fibroblast growth factor-2 (FGF2; R&D Systems) and 20 ng/ml PDGF-AA (R&D Systems). For differentiation into oligodendrocytes, OPCs were plated in neural basal medium plus 0.4 ng/ml 3,3′,5-triiodothronine (T3; MilliporeSigma), 100 ng/ml Noggin, 10 μM dibutyryl cyclic-AMP sodium salt (cAMP; MilliporeSigma), 100 ng/ml insulin-like growth factor-1 (IGF-1; R&D Systems), and 10 ng/ml neurotrophin-3 (NT-3; R&D Systems). Cultures were maintained in 37°C with 5% CO₂ supplemented with 50U penicillin-streptomycin.

Primary mouse OPC culture

Wild-type, mixed sex B6CBACaF1/J-Aw-J/A mice (The Jackson Laboratory) were euthanized two days after birth and cerebral cortex was dissected in ice cold DMEM/F12 with 50U penicillin-streptomycin. The meninges were removed and tissue was dissociated to single cells using the Tumor Dissociation Kit (Miltenyi Biotec) with gentle trituration. Cell aggregates were excluded using a 70 mm Corning Falcon cell strainer (Thermo Fisher Scientific). Cells were cultured as described in mouse OPC culture. Mice were handled under protocols approved by Case Western Reserve University School of Medicine’s IACUC.

hOPC culture

Human OPCs were generated from iPSCsas previously described (Nevin et al., 2017) from a 16 year old male donor. For seven days, cells were maintained in DMEM/F12, 1X high insulin N-2 supplement, 10 μM SB431542, 250 nM LDN189193 (R&D Systems), and 100nM ATRA. For the next four days, cells were fed daily with DMEM/F12, 1X low insulin N-2 supplement (Thermo Fisher Scientific), 100 nM ATRA, and 1 μM smoothened agonist (SAG; MilliporeSigma). Cells were then lifted with a cell scraper and plated into ultra-low attachment plates (Corning Life Sciences) for eight days during which two-thirds medium was replaced with DMEM/F12, 1X low insulin N-2 supplement, 1X B-27 supplement without vitamin A, 100 nM ATRA, and 1 μM SAG every other day. For the next 10 days, cells underwent two-thirds medium changes of DMEM/F12, 1X low insulin N-2, 1X B-27, 10 ng/mL platelet-derived growth factor (PDGF-AA; R&D Systems), 10 ng/mL IGF-1, 5 ng/mL hepatocyte growth factor (HGF; R&D Systems), 10 ng/mL NT-3, 60 ng/mL T3, 100 ng/mL biotin, 1 μM cAMP, and 25 μg/mL insulin. Cells were then plated onto poly-L-ornithine- and laminin-coated plates for 30 days during which two-thirds medium was replaced with DMEM/F12, 1X low insulin N-2, 1X B-27, 10 ng/mL PDGF-AA, 10 ng/mL IGF-1, 5 ng/mL HGF, 10 ng/mL NT-3, 60 ng/mL T3, 100 ng/mL biotin, 1 μM cAMP, and 25 μg/mL insulin every other day. OPC cultures were maintained in PDGF-AA, IGF-1, and HGF containing media through day 70. For OPC differentiation into oligodendrocytes, cells were passaged into medium lacking PDGF-AA, IGF-1, and HGF. All medium was supplemented with 5 U/ml penicillin-streptomycin and cells were maintained at 37°C with 5% CO₂.

DRG co-culture

DRG cultures were grown as previously described (Nevin et al., 2017). Briefly, 5x10⁴ DRG neurons derived from E15.5 embryos were plated into 24-well VisiPlates (PerkinElmer) precoated with rat tail collagen (Thermo Fisher Scientific) and dried. Cells were maintained in Minimum Essential Media (MEM; Thermo Fisher Scientific), 10% FBS, 2% glucose (MilliporeSigma), plus 100 ng/ml NGF (R&D Systems) for 24 h and then fed with MEM, 2% glucose, 1X high insulin N-2 supplement, 245 ng/ml FDU (MilliporeSigma), 245 ng/ml uridine (MilliporeSigma) and 100 ng/ml NGF. Cells were maintained until day 20 (human OPCs) or day 30 (mouse OPCs) prior to seeding 5x10⁴ OPCs per well. Cultures were supplemented with 5 U/ml penicillin-streptomycin and maintained at 37°C with 5% CO₂. Plates were fixed at day 37 or 40 and immunostained as described below. Rats were handled under protocols approved by Case Western Reserve University School of Medicine’s IACUC.

Multiple sclerosis patient tissue

Human brains (male and female ages 27 – 76; Table S5) were collected as part of the tissue procurement program approved by the Cleveland Clinic Institutional Review Board. All experiments were carried out in accordance with the relevant Cleveland Clinic Institutional regulations and guidelines.

Multiple sclerosis genotyped samples

Multiple sclerosis genetic data was obtained from Wellcome Trust Case Control Consortium. This includes 9,772 cases and 5,175 controls of European descent. Prior to imputation, genotypes and samples were filtered based on the quality control detailed by WTCCC2 (Sawceret al., 2011). IMPUTE2 with the 1000 genomes integrated reference panel (Phase 1) was utilized to perform imputation (Howie et al., 2009). SNPs departing from Hardy Weinberg Equilibrium (HWE p < 1E-6) were excluded. Two filters were applied to identify high quality SNPs. First, we required SNPs have INFO scores > 0.4 in > 95% of samples. Second, we excluded genotypes that did not surpass a stringent imputation filter requiring > 0.9 confidence for the genotype call and < 0.3 for the remaining two potential genotypes for at least 10% of cases and 10% controls.

METHOD DETAILS

Hi-C data analysis

Preprocessed promoter capture Hi-C data was obtained for 15 immune cell types. Physical interactions with promoters were identified using the CHiCAGO pipeline (Cairns et al., 2016) as previously described (Javierre et al., 2016). Interactions with a CHiCAGO score ≥ 5 in at least one sample were utilized for subsequent analyses. Previously published NPC Hi-C data (Dixon et al., 2015) were incorporated to capture potential CNS specific interactions. NPC Hi-C data was processed with HiC-Pro (Servant et al., 2015) and Bowtie2 (Langmead and Salzberg, 2012) with bin size 20-kb. Pairwise contacts with ICE normalized scores > 4 (top 10^th percentile of intrachromosome contacts), were identified as putative physical interactions. Pairwise physical interactions were filtered to include only those that involved a transcription start site (TSS) in at least one end.

Curation of ChIP-seq panel

Data acquisition, processing and QC

A panel of 20 cell types from the immune and central nervous system were acquired from previously published ChIP-seq studies. The majority of immune cell ChIP-seq experiments were acquired from the BLUEPRINT consortium (Adams et al., 2012). BLUEPRINT preprocessed data, including replicates for each cell type where available, were downloaded. Published BLUEPRINT samples met standard quality control metrics including fraction of reads in peaks (FriP) score > 0.01, normalized strand correlation (NSC) > 1.05, and relative strand correlation (RSC) > 0.8.

Previously published H3K27ac ChIP-seq for Th17 (Aschenbrenner et al., 2018), (GABA)ergic inhibitory neurons, glutamatergic excitatory neurons and oligodendrocytes (Kozlenkov et al., 2018) were also obtained. These samples were evaluated with ChIPQC (Carroll et al., 2014) each were found to pass the same ChIP QC metrics as BLUEPRINT (FriP score > 0.01, NSC > 1.05 and RSC > 0.8). We processed these raw datasets following a pipeline similar to BLUEPRINT (Chen et al., 2016) on these additional samples, detailed below. We aligned reads to hg38 (GRCh38) using BWA quality threshold of 15 (bwa aln -q 15). Duplicate reads and reads with mapping quality < 15 were discarded using SAMtools (Li et al., 2009). Peaks were called using MACS2 (Zhang et al., 2008) standard parameters with narrow peaks option. Fragment size input for MAC2 algorithm was calculated using PhantomPeakQualTools (Marinov et al., 2014).

As an additional QC metric, hierarchical clustering was utilized to compare curated samples including sample replicates. We verified that immune cell types clustered as expected based on cellular lineage as well as correlated well within replicates (Figure S2). Finally, all of the samples that passed these standard QC metrics were visually evaluated on the genome browser. The replicate with the least background was selected to complete our final panel of 20. (Table S1).

Consensus peak list

All MACS2 narrow peaks with p value < 1E-9 were identified for the panel of 20 cell types. Peaks were extended by 300bp in either direction and then collapsed to create a consensus peak list used in subsequent analyses below.

Outside variant approach

Definition of candidate outside variants

Candidate outside variants were identified using ChIP-seq and Hi-C data from immune and neural cell types and further filtered based on genotype frequency and LD as detailed below (Figure S1). Previously published MS risk SNPs (n = 163) were identified from the GWAS catalog (2/11/2018; (Buniello et al., 2019). GWAS SNPS and SNPs in LD with the lead SNP (r² > 0.8) were compared to promoter-centric Hi-C interactions identified in 15 immune cell types and NPCs (see Hi-C data analysis). Putative gene targets were defined as any promoter that physically associated with the GWAS SNPs (or LD SNPs r² > 0.8) in at least one cell type. All SNPs in Hi-C interactions with these putative gene targets were identified to define putative candidate outside variants. Candidate outside variants were filtered to include only those those that were in sites enriched for H3K27ac in at least one of the 20 cell types included in our panel, defined by our consensus peak list (Curation of ChIP-seq panel).

This resulted in a list of candidate outside variants (n = 54,471 paired to a GWAS SNP based on shared physical interaction with the same target promoter. The candidate outside variant list was further filtered based on genotype frequency and LD with the GWAS SNP (described below). Nine possible genotype combinations are produced from each GWAS SNP + outside variant pair. The frequency of each genotype combination in the MS GWAS panel was quantified. Genotype combinations that were < 1% of cases or controls were excluded.

Additionally, only outside variants for which at least 6 of the 9 genotype combinations surpassed this frequency were considered. This filter enables exclusion of variants that are in strong LD with the GWAS SNP (note: if the GWAS SNP and outside variant were in perfect LD only 3 genotype combinations would be observed). After filtering, 48,455 outside variants remained across 126 GWAS loci. On average 560 outside variants were tested per GWAS SNP (range: 27-1,804 outside variants per GWAS SNP).

Identifying outside variants that alter risk

The impact of outside variants on clinical risk was evaluated as previously described (Corradin et al., 2016). First, individuals were stratified based on the genotype of the GWAS SNPs and then again by the genotype of the outside variant. This resulted in up to 9 possible genotype combinations (Figure S1). Risk was calculated for each of these groups separately. To evaluate whether the change in odds ratio was significant, the outside variant was randomly permuted such that outside variant allele frequency remains constant and the impact of the GWAS SNP on risk remained unchanged. For example, if all individuals homozygous for the risk allele had an odds ratio of 1.1, permuting the outside variant (i.e., randomly segregating risk/risk individuals) would generate a null distribution of odds ratios with a median of approximately 1.1. This approach enables quantification of whether the observed impact of the outside variant genotype is more than expected by random chance. Each genotype combination was evaluated independently such that up to nine p values are generated per outside variant GWAS pair. This approach enables identification of outside variants that alter clinical risk beyond the effect of the primary GWAS SNP while being agnostic to the type of interaction between GWAS SNP and outside SNP (i.e., additive, epistatic, dominant interactions modes can all be identified using this method Figure S1).

Adaptive permutations

An adaptive permutation approach was utilized to estimate p values and reduce the substantial computational burden of permutation-based tests. We calculated the 95% confidence interval for each p value iteratively at multiple steps through the permutation process. If the upper bound of the 95% confidence interval was below the p value threshold the genotype was defined a significant and no additional permutation were performed. If the lower bound of the confidence interval is above the p value threshold the genotype was defined as not significant. For all others, additional permutations were performed up to a maximum of 10 million permutations.

Multi-test correction

As each genotype combination is evaluated independently, multi-test correction accounted for the number of outside variants tested for each GWAS SNP and the total number of genotypes evaluated for each outside variant. For example, if all 9 genotype combinations were tested for 1,500 GWAS SNP-outside variants pairs then the p value threshold would be 0.05/(9*1500 = 3.7E-6). The most stringent p value threshold was 3.9E-6

Pathogenic cell type prediction

Overview

Outside variant results described above were incorporated with H3K27ac ChIP-seq data to identify likely pathogenic cell types for each GWAS risk locus. Outside variant results were segmented utilizing the ChIP-seq peaks from the consensus peak list as boundaries. The set of outside variants paired to each GWAS locus were evaluated separately. For each GWAS locus, all ChIP peaks containing at least one outside variant that significantly altered risk were defined as ‘sites that alter risk’ and peaks for which at least one outside variant was tested, but none were significant were classified as ‘sites that do not alter risk’ (Figure S4). This strategy prevents overrepresentation of genomic regions with multiple outside variants. Additionally, it reduces the disproportional weighing of enhancer regions due to differences in LD without excluding the contribution of multiple functional variants in LD (as has been previously described (Canver et al., 2017; Corradin et al., 2014; Fortini et al., 2014; Glubb et al., 2015; Guo et al., 2015; He et al., 2015; Rajagopal et al., 2016)

Quantifying ChIP-seq differences

H3K27ac ChIP-seq signal was quantified for each peak region in order to compare sites that alter risk to those that do not. ChIP signal was quantified in 40 equal sized bins ± 1500bp from the center of each peak from the consensus peak list. The median signal was extracted from wig files with step size of 10bp for each bin. The resulting matrix was then z-scored. For each GWAS locus, the signal across all sites that alter risk was aggregated by taking the average of each bin across the set of peaks. The average signal in each bin was also aggregated across sites that do not alter risk. One-sided Kruskal-Wallis tests were performed to identify cell types in which H3K27ac signal at sites that alter risk was greater than the signal at sites that do not alter risk. This process was repeated for each cell type of interest. P values were multi-test corrected for the total number of GWAS SNPs (n = 126) and total number of cell type evaluated (n = 20) resulting in a threshold of p < 2E-5.

LD score regression

Overview

LDSR was utilized to estimate heritability enrichment across annotated SNP lists for (1) H3K27ac enriched sites found in each of the 20 cells in our panel (Figure 1C), (2) clusters of GWAS loci defined by outside variant cell type predictions (Figure 1D), and (3) transcriptional pausing pathway gene sets (Figure 2G). For each, European LD scores and weights were downloaded from the LDSC (Bulik-Sullivan et al., 2015). Formatted summary statistics for multiple sclerosis and 27 additional traits were obtained https://data.broadinstitute.org/alkesgroup/sumstats_formatted/. The GWA studies utilized are listed in Table S1. For each SNP list of interest (detailed below) annotation-specific LD scores were calculated as recommended https://github.com/bulik/ldsc/wiki/ using HapMap3 SNPs and 1000 genomes phase 3 reference files (Auton et al., 2015). Heritability enrichment calculations were conditioned on the recommended ‘baseline model’, comprised of multiple functional categories including conserved regions and multiple enhancer classifications.

Annotated SNP lists

Annotated SNP lists were defined as follows:

H3K27ac global enrichment

To estimate MS heritability enrichment in H3K27ac enriched sites found in each cell type, SNPs within ChIP-seq peaks identified by MACS2 with p < 1E-9 were utilized.

Outside variant defined clusters

LDSR was also used to evaluate heritability enrichment across multiple diseases for the GWAS loci grouped by cell type prediction. SNPs within 250kb of each GWAS SNP within the outside variant defined cluster were utilized. The significance threshold was defined through Bonferroni multi-test correction (8 groups of GWAS loci, 28 disorders = 224 tests).

Transcriptional pausing pathway

Two sets of gene list were generated to evaluate heritability enrichment of transcriptional elongation pathway. The first came from Reactome (Fabregat et al., 2018). Gene sets and their interactors (IntAct database; (Hermjakob et al., 2004); confidence score > 0.45) involved in pre-transcription events (ID:R-HSA-674695.3), transcription pre-initiation and promoter opening (ID:R-HSA-73779.2), transcription initiation and promoter clearance (ID: R-HSA-76042.3), RNAP II CTD phosphorylation and interaction with CE (ID:R-HSA-77075.2), and transcriptional elongation (ID:R-HSA-75955.2) were downloaded from Reactome. SNPs within 500kb of the TSSs of these gene sets were utilized. The second gene list was curated based on the gene ontology GO-0006368, “Transcription elongation from RNA polymerase II promoter” (Ashburner et al., 2000). This gene ontology adapted to include missing critical components for pause release regulation such as the 7SK complex which includes HEXIM1/2, LARP7 and RN7SK. These genes were supplemented to GO-0006368 and components involved in the assembly of RNAPII such as GTF2A1 were excluded. Each gene and the reason for its inclusion is listed in Table S6. SNPs within 500kb of the TSSs of these gene sets were utilized.

CRISPR/Cas9 outside variants experiments

Outside variant selection and guide RNA design

Outside variants that significantly alter risk for the HEXIM1, BRD3, and SOCS1 loci with robust H3K4me1/H3K27ac enrichment were selected. Single guide (sg) RNAs were designed using the CRISPOR online tool (Haeussler et al., 2016)(Table S2). sgRNA oligos (Integrated DNA Technologies, Coralville, IA) were cloned into pSpCas9(BB)-2A-GFP vector (PX458, Addgene plasmid #48138, Watertown, MA (Ran et al., 2013).

Transfection and sorting

Cells were electroporated using the Neon Transfection System (Thermo Fisher Scientific). 30 μg of plasmid DNA and 2x10® cells were used per transfection for Jurkat cells, 20 μg plasmid DNA and 2x10® cells were used for GM12878 cells, and 30 μg plasmid DNA and 2.5x10® cells for NPCs. Jurkat and NPC parameters were: 1350 V, 10 ms, 3 pulses. GM12878 parameters were: 1150 V, 30 ms, 2 pulses. After transfection, Jurkat cells and GM12878 cells were incubated for 48 hours in maintenance media lacking penicillin-streptomycin. NPCs were incubated at 37°C at 5% CO₂ for 24 hours in supplemented maintenance media plus ROCK inhibitor lacking penicillin-streptomycin followed by 48 hours in supplemented maintenance media with penicillin-streptomycin after transfection. GFP+ cells were subsequently sorted via FACSAria II (BD Biosciences, San Jose, CA).

Genomic DNA isolation and mutation validation

Genomic DNA from at least 5x10⁴ GFP+ cells was isolated and amplified using the GeneArt Genomic Cleavage Detection Kit (Thermo Fisher Scientific; see Table S2 for primer sequences). Cleavage assays were performed as per the manufacturer’s protocol. Images were acquired using the Bio-Rad Gel Doc XR+ system and the Image Lab software. Non-overlapping PCR products were pooled for sequencing. CRISPResso tool was utilized to analyze indels generated by select sgRNAs (Pinello et al., 2016).

RNA isolation and qRT-PCR

RNA was isolated from at least 3x10⁵ GFP+ Jurkat cells, GM12878 cells, and NPCs using the miRNeasy Mini Kit (QIAGEN, Hilden, Germany) and cDNA was generated using Superscript IV First-Strand Synthesis System (Thermo Fisher Scientific) per the manufacturers’ protocols. qPCR was performed using TaqMan Universal PCR Master Mix (Thermo Fisher Scientific) with TaqMan probes on the QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific). Fold change was determined using the AAC_T method, using samples transfected with Cas9 plasmid without a guide RNA as a comparator. Samples were run in technical triplicate or quadruplicate and the average result from each biological replicate is reported. Data was analyzed and visualized using GraphPad Prism 7.0.

Screening

Poly-D-lysine coated 384-well CellCarrier ultra plates (PerkinElmer) were coated with laminin (4 μg/mL). OPCs were plated at 1.25x10⁴ per well in differentiation media (DMEM/F12; 1X N-2 high insulin supplement; 1X B-27 without vitamin A; 1X GlutaMAX; 100 ng/ml Noggin; 10 ng/ml NT-3; 50 μM cAMP; and 100 ng/ml IGF-1). Cells were left at room temperature for 45 minutes to attach before addition of drug. A library of 3,141 bioactive small molecules with well-an notated targets was added to the cells (columns 1-22, Figures S7C and S7D) by pin transfer at a final concentration of 2 μM, and cells were incubated at 37°C for 1 hour. 20 ng/ml FGF was added to negative control wells (column 24, Figures S7C and S7D) and 40 ng/ml T3 to all other wells. T3 alone was used as a positive control (column 23, Figures S7C and S7D). Cells were fixed and stained for oligodendrocyte lineage marker OLIG2 (MilliporeSigma), oligodendrocyte marker MBP (Abeam), and DAPI (MilliporeSigma) after 72 hours.

A total of 10 plates were utilized to assess the 3,141 compound library. All plates were processed simultaneously to reduce variability. Five 20X fields per well were captured using an Operetta high content imaging system (PerkinElmer). Harmony software (PerkinElmer) was used to identify nuclei using DAPI. Cells with nuclear OLIG2 staining and those with peri-nuclear MBP staining were counted in each well and percent positivity of each calculated. The percent MBP positive cells in each treatment was normalized to controls on a plate-by-plate basis. Compounds that were in the bottom 10% of the library for either total cell count or percent oligodendrocyte lineage (OLIG2+) cells were excluded. The median and standard deviation for each well position was calculated across the 10 plates to assess for well position effects as shown in Figure S7. The standard deviation across different compounds in the same well position was similar across the plates. Some well-to-well variation in median MBP positive percentages was observed which could contribute to the false negative rate. The two percent of compounds that produced the smallest percent oligodendrocytes (MBP+) after these exclusions were selected for further analysis.

Immunostaining

Cells were fixed for 15 minutes in fresh cold 4% paraformaldehyde (Electron Microscopy Sciences), and permeabilized for 10 minutes with 0.2% Triton X-100 (MilliporeSigma) in pH 7.4 PBS. Blocking was performed for 1 hour with 10% normal donkey serum (Jackson ImmunoResearch). Cells were incubated in primary antibody diluted in blocking solution overnight at 4°C, then washed with PBS. Alexa Fluor secondary antibodies (anti-rabbit Alexa Fluor 594 and anti-rat Alexa Fluor 488; 1:500; Thermo Fisher Scientific) were diluted in blocking solution and incubated on cells for 1 hour at room temperature. Nuclei were stained with DAPI in PBS for 5 minutes and then washed with PBS. Primary antibodies used were as follows: Sox10 (R&D Systems; 1:100), OLIG2 (MilliporeSigma; 1:1000), NF (Covance; 1:1000).

RNA-seq

RNA from 1x10⁶ cells was extracted using TRIzol Reagent (Thermo Fisher Scientific) followed by phase separation using Phase Lock tubes (5 Prime) and isolation with the miRNeasy kit (QIAGEN). Libraries were prepared using the TruSeq Stranded Total RNA kit (Il-lumina) and sequenced on Illumina instruments at the Case Western Reserve University Genomics Core. Reads were aligned to the iGenomes mm9 build using Tophat (Trapnell et al., 2010). FPKMs were calculated using Cufflinks (Trapnell et al., 2010) then quantile normalized. FPKMs were floored to allow log transformations by replacing all values below FPKM = 0.25 with 0.25. Differential expression testing was performed using Cuffdiff (Trapnell et al., 2010).

RNAP II ChIP-seq

Experiments and analysis were carried out as described(Factor et al., 2014; Schmidt et al., 2009), except that 5x10⁷ cells were used for RNAP II ChIP, and 10 mg of 8wg16 antibody was used (Abcam). Briefly, cells were crosslinked with 11% buffered formaldehyde (MilliporeSigma) and sonicated. Antibodies were incubated with Dynabeads Protein G (Thermo Fisher Scientific) before proceeding to immunoprecipitation. ChIP-seq libraries were sequenced on Illumina instruments at the Case Western Reserve University Genomics Core Facility. Adaptor sequence and quality trimming were performed using the FASTX-Toolkit. Reads were aligned to iGenomes mm9 builds using Bowtie (Langmead and Salzberg, 2012). Duplicate reads were removed using SAMtools (Li et al., 2009) and peaks called using MACS using input DNA as control. Wiggle tracks generated by MACS were converted to BigWig using the UCSC Genome Browser’s wigToBigWig tool and visualized on the UCSC Genome Browser. Gene body read counts were generated for the region from 500bp downstream of the transcription start site to the transcription termination site using the annotatePeaks tool in the HOMER package (Heinz et al., 2010).

Hexim1 expression and OPC maturation

Generation of Hexim1 CRISPRa/CRISPRi OPCs

CRISPR activation (CRISPRa) and CRISPR interference (CRISPRi) OPCs were generated following a similar protocol used previously (Hubler et al., 2018; Tripathi et al., 2019). Guide sequences for CRISPRa and CRISPRi were curated from the CRISPRa v2 and CRISPRi v2 libraries, respectively (Horlbecket al., 2016). For CRISPRa, guides were cloned into the lentiSAMv2 backbone (Addgene plasmid #75112; (Joung et al., 2017). The activation helper plasmid lentiMPHv2 (Addgene plasmid #89308; (Joung et al., 2017) was also co-transduced for CRISPRa OPCs. Guides for CRISPRi were cloned into the pLV hU6-sgRNA hUbC-dCas9-KRAB-T2a-Puro backbone (Addgene plasmid #71236; (Joung et al., 2017). Specific guides used in experiments are detailed in Table S3. Plasmids used were confirmed to have the correct inserted oligomer using Sanger sequencing. Lentivirus was produced using Lenti-X Packaging Single Shots following the manufacturer’s protocol (Takara Bio, Mountain View, CA). Lentiviral media was collected after 3 days, filtered, supplemented with OPC growth factors PDGF and FGF, and added to OPCs at a ratio of 1:2 viral media to growth media. After 24 hours of incubation with viral media, transduced cells were switched to fresh OPC growth media and allowed to recover for 48 hours. CRISPRa and NTC OPCs were then selected with a lethal dose of blasticidin (10 mg/mL; Thermo Fisher Scientific) and hygromycin (100 mg/ml; Thermo Fisher Scientific) for 96 hours. CRISPRi OPCs were selected by switching to OPC growth media supplemented with a lethal dose of puromycin (500 ng/ml; Thermo Fisher Scientific) for 96 hours. OPCs were then allowed to recover for at least 24 hours following removal of selection and frozen down in aliquots for future use. For all experiments, the CRISPRa/i targeting and non-targeting OPCs were derived from the same batch of OPCs and infected and selected simultaneously.

RNA isolation and qRT-PCR

qRT-PCR was performed to validate a reduction or overexpression of gene targets of interest compared to the non-targeting control. At least 10⁶ OPCs were lysed in TRIzol reagent followed by phenol-chloroform extraction and processing with the RNeasy Mini Kit (QIAGEN). cDNA was then generated using the iSCRIPT kit following the manufacturer’s instructions (Bio-Rad). qRT-PCR was performed using pre-designed TaqMan gene expression assays (Table S3; Thermo Fisher Scientific). qPCR was performed using the Applied Biosystems 7300 Real-Time PCR System and probes were normalized to the Rpl13a endogenous control.

MS patient microarray

HEXIM1/2 and BRD3 expression was analyzed using previously published microarray data (Huynh et al., 2014). Here, unidentified post-mortem brain frontal lobe specimens were obtained from the UK Multiple Sclerosis Tissue Bank. White matter regions from 13 controls and 12 MS patient lesions were evaluated. RMA normalization was utilized to quantify gene expression.

MS patient tissue

A rapid autopsy protocol was used to remove the brains within 6 hours post mortem. The brains were then sliced 1 cm thick and alternating sections were frozen or fixed in 4% paraformaldehyde. Frozen 30 mm sections were cut for characterization by immunohistochemistry and for assessment of demyelination and other pathology using previously described protocols (Dutta et al., 2006).

Analysis of Hexim1 in MS patient tissue

Patient demographics are listed in Table S4. Sections were washed 3 times in PBS and then microwaved twice in 10 mM citric acid buffer (pH 6.0; MilliporeSigma) for 5 minutes and allowed to cool to room temperature. The sections were then permeabilized with 2% Triton X-100 in PBS for 30 minutes and then incubated with antibodies against HEXIM1 (Abcam) and PLP (produced in house at the Cleveland Clinic hybridoma facility) for 3 days followed by fluorescently conjugated secondary antibodies (anti-rabbit Alexa Fluor 594 and anti-rat Alexa Fluor 488) for 2 hours. PLP1 staining was utilized to identify oligodendrocyte lineage cells.

Atotal of 5 MS patients (2 primary progressive and 3 secondary progressive) and 3 controls were assessed (Table S6). MS patient lesions were randomly sampled in multiple regions of the forebrain. The number of HEXIM1+ cells counted from multiple 0.07 mm² areas from the samples are plotted in Figure 5C.

QUANTIFICATION AND STATISTICAL ANALYSES

Statistical details can be found in figure legends.

DATA AND CODE AVAILABILITY

https://github.com/corradin-lab/outside-variants

The accession number for the ChIP and RNA-seq data generated for this study is GEO: GSE142084 and GSE142085

Supplementary Material

NIHMS1581397-supplement-1.pdf^{(789.1KB, pdf)}

NIHMS1581397-supplement-2.pdf^{(1MB, pdf)}

NIHMS1581397-supplement-3.pdf^{(2.1MB, pdf)}

NIHMS1581397-supplement-4.pdf^{(27MB, pdf)}

NIHMS1581397-supplement-5.pdf^{(1.2MB, pdf)}

NIHMS1581397-supplement-6.xlsx^{(49.2KB, xlsx)}

NIHMS1581397-supplement-7.xlsx^{(52.7KB, xlsx)}

NIHMS1581397-supplement-8.pdf^{(2.5MB, pdf)}

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
8WG16	Abcam	Cat#ab817; RRID:AB_306327
Donkey anti-rabbit AF594	Thermo Fisher Scientific	Cat#A-21207; RRID:AB_141637
Donkey anti-rat AF488	Thermo Fisher Scientific	Cat#A-21208; RRID:AB_141709
HEXIM1	Abcam	Cat#ab25388; RRID:AB_2233058
MBP	Abcam	Cat#ab7349; RRID:AB_305869
NF	Covance	Cat#SMI311; RRID:AB_509991
NF	Covance	Cat#SMI312; RRID:AB_2314906
OLIG2	MilliporeSigma	Cat#AB9610; RRID:AB_570666
PLP	Cleveland Clinic Hybridoma Core Facility	AA3
SOX10	R&D Systems	Cat#AF2864; RRID:AB_442208
Biological Samples
Human, female and male brain tissue (ages 27 – 76)	Cleveland Clinic Mellen Center for Multiple Sclerosis	Table S5
Chemicals, Peptides, and Recombinant Proteins
2-mercaptoethanol	MilliporeSigma	Cat#M3148
AlbuMAX I Lipid-Rich Bovine Serum Albumin	Thermo Fisher Scientific	Cat#11020021
Ascorbic Acid	MilliporeSigma	Cat#PHR1008
ATRA	MilliporeSigma	Cat#R2625
B-27 without vitamin A	Thermo Fisher Scientific	Cat#12587
Biotin	AnaSpec	Cat#AS-21100
Blasticidin	Thermo Fisher Scientific	Cat#R21001
cAMP	MilliporeSigma	Cat#D0260
Citric acid buffer	MilliporeSigma	Cat#21545
Collagenase	Thermo Fisher Scientific	Cat#17104019
DAPI	MilliporeSigma	Cat#D8417
DL-Lactic acid syrup	MilliporeSigma	Cat#L1250
DMEM/F12	Thermo Fisher Scientific	Cat#11320-033
Dorsomorphin (NPC generation)	Thermo Fisher Scientific	Cat#309310
Dorsomorphin (OPC generation)	MilliporeSigma	Cat#P5499
Dynabeads Protein G	Thermo Fisher Scientific	Cat#10003D
FDU	MilliporeSigma	Cat#F0503
Fetal bovine serum	ATCC	Cat#30-2020
FGF2	R&D Systems	Cat#233-FB-25
Formaldehyde	MilliporeSigma	Cat#818708
Gem21 Neuroplex w/o vitamin A	Gemini Bio-Products	Cat#400-161
Glucose	MilliporeSigma	Cat#G7528
GlutaMax	Thermo Fisher Scientific	Cat#35050079
HGF	R&D Systems	Cat#294-HG
High-insulin N-2 supplement	R&D Systems	Cat#AR009
Human insulin	MilliporeSigma	Cat#I9278
Human recombinant FGF-Basic (AA 1-155)	Thermo Fisher Scientific	Cat#PHG0264
Hygromycin	Thermo Fisher Scientific	Cat#10687010
IGF-1	R&D Systems	Cat#291-G1-200
Knockout DMEM	Thermo Fisher Scientific	Cat#10829018
Knockout Serum Replacement	Thermo Fisher Scientific	Cat#10828028
L-glutamine	GE Healthcare Life Sciences	Cat#82007-322
Laminin	MilliporeSigma	Cat#L2020
LDN189193	R&D Systems	Cat#04-0074
Lenti-X Packaging Single Shots	Takara Bio	Cat#631278
Low-insulin N-2 supplement	Thermo Fisher Scientific	Cat#17502001
Matrigel	Corning Life Sciences	Cat#356231
MEM	Thermo Fisher Scientific	Cat#11095080
N2 NeuroPlex Serum-Free supplement	Gemini Bio-Products	Cat#400-163
NaCl	Thermo Fisher Scientific	Cat#AM9760G
NaOH	MilliporeSigma	Cat#109137
Neurobasal medium	Thermo Fisher Scientific	Cat#21103049
NGF	R&D Systems	Cat#556-NG
Noggin	R&D Systems	Cat#3344-NG
Nonessential amino acids	Thermo Fisher Scientific	Cat#11140050
Normal donkey serum	Jackson ImmunoResearch	Cat#017-000-121
Normal donkey serum	Jackson ImmunoResearch	RRID:AB_2337258
NT-3	R&D Systems	Cat#267-N
Paraformaldehyde	Electron Microscopy Sciences	Cat#15710
PDGF-AA	R&D Systems	Cat#221-AA-050
Penicillin-streptomycin	Thermo Fisher Scientific	Cat#15140122
Poly-L-ornithine	MilliporeSigma	Cat#P3655
Puromycin	Thermo Fisher Scientific	Cat#A1113802
Rat tail collagen	Thermo Fisher Scientific	Cat#A1048301
ROCK Inhibitor Y27632	MilliporeSigma	Cat#SCM075
RPMI 1640	Thermo Fisher Scientific	Cat#11875119
SAG	MilliporeSigma	Cat#566660
SB431542	MilliporeSigma	Cat#S4317
Sodium pyruvate	Thermo Fisher Scientific	Cat#11360070
StemPro Accutase Cell Dissociation Reagent	Thermo Fisher Scientific	Cat#A1110501
T3	MilliporeSigma	Cat#ST2877
Triton X-100	MilliporeSigma	Cat#1086431000
TRIzol Reagent	Thermo Fisher Scientific	Cat#15596018
Uridine	MilliporeSigma	Cat#U3003
Critical Commercial Assays
GeneArt Genomic Cleavage Detection Kit	Thermo Fisher Scientific	Cat#A24372
iScript cDNA Synthesis Kit	Bio-rad	Cat#1708891
SuperScript IV First-Strand Synthesis System	Thermo Fisher Scientific	Cat#18091200
miRNeasy Mini Kit	QIAGEN	Cat#217004
RNeasy Mini Kit	QIAGEN	Cat#74104
TaqMan Universal PCR Master Mix	Thermo Fisher Scientific	Cat#4304437
TruSeq Stranded Total RNA Kit	Illumina	Cat#20020596
Tumor Dissociation Kit	Miltenyi Biotec	Cat#130-095-929
Deposited Data
CNS H3K27ac ChIP-seq data	Kozlenkov et al., 2018	N/A
CRISPRa/i v2 libraries	Horlbeck et al., 2016	N/A
BLUEPRINT	Adams et al., 2012	http://www.blueprint-epigenome.eu RRID:SCR_003844
European LD scores	Auton et al., 2015	https://www.internationalgenome.org RRID:SCR_008801 https://data.broadinstitute.org/alkesgroup/LDSCORE/
Gene Ontology	Ashburner et al., 2000; The Gene Ontology Consortium, 2019	http://geneontology.org RRID:SCR_002811
GWAS Catalog	Buniello et al., 2019	https://www.ebi.ac.uk/gwas/ RRID:SCR_012745
GWAS Summary Statistics	Price lab	https://data.broadinstitute.org/alkesgroup/sumstats_formatted/
HapMap3	Wellcome Sanger Institute	https://www.sanger.ac.uk/resources/downloads/human/hapmap3.html RRID:SCR_004563
Homo sapien hg38 genome build	NCBI; Genome Reference Consortium	https://www.ncbi.nlm.nih.gov/assembly/GCF_000001405.26/
IntAct database	Hermjakob et al., 2004	https://www.ebi.ac.uk/intact/ RRID:SCR_006944
MS patient microarray data	Huynh et al., 2014	N/A
Mus musculus mm9 genome build	NCBI; Mouse Genome Sequencing Consortium	https://www.ncbi.nlm.nih.gov/assembly/GCF_000001635.18/
NPC Hi-C	Dixon et al., 2015	N/A
Primary brain Hi-C data	Jung et al., 2019	N/A
Reactome	Fabregat et al., 2018	https://reactome.org RRID:SCR_003485
RNAP II ChIP-seq H3K27ac ChIP-seq	This paper	GSE142084
RNA-seq	This paper	GSE142085
Th17 H3K27ac ChIP-seq data	Aschenbrenner et al., 2018	N/A
Experimental Models: Cell Lines
Human, Female: GM12878 LCL cell line	Coriell Institute	Cat#GM12878; RRID:CVCL_7526
Human, Male: Jurkat Cell Line, Clone E6-1	ATCC	Cat#TIB-152; RRID:CVCL_0367
Human, Female blastocyst: WIBR3 ESCs	Lengner et al., 2010	RRID:CVCL_9767
Human, 88 year old Female and 29 year old Male: Fibroblasts for iPSC and NPC generation	Coriell Institute	Cat#AG07657 and AG07599 RRID:CVCL_4P27, CVCL_4N84
Human, 16 year old male: Fibroblast for iPSC and oligodendrocyte generation	This paper
Experimental Models: organisms/Strains
Mouse, mixed sex, 2 days postnatal: B6CBACaF1/J-Aw-J/A	Jackson Laboratory	Stock#001201 RRID:IMSR_JAX:001201
Mouse, Male: 129S/SvEv E3.5 for EpiSC isolation	Jackson Laboratory	MGI:3050593 RRID:MGI:3050593
Mouse, Female and Male: Embryonic fibroblasts isolated from E13.5 embryos of CD-1 mice	Charles River Laboratories	Strain code#022 RRID_IMSR_CRL:022
Sprague Dawley Rat, not sexed: E15.5 embryos for DRG culture	Charles River Laboratories	Strain code#400
Oligonucleotides
Primers	This manuscript	Table S3
sgRNAs	This manuscript	Table S3
Taqman Probes	Thermo Fisher Scientific	Table S3
Recombinant DNA
lentiMPHv2	Joung et al., 2017	Addgene plasmid #89308 RRID:Addgene_89308
lentiSAMv2	Joung et al., 2017	Addgene plasmid #75112 RRID:Addgene_75112
pLV hU6-sgRNA hUbC-dCas9-KRAB-T2a-Puro	Joung et al., 2017	Addgene plasmid #71236 RRID:Addgene_71236
pSpCas9(BB)-2A-GFP vector (PX458)	Ran et al., 2013	Addgene plasmid #48138; RRID:Addgene_48138
Software and Algorithms
Bowtie2	Langmead and Salzberg, 2012	http://bowtie-bio.sourceforge.net/bowtie2/index.shtml RRID:SCR_005476
CHiCAGO	Cairns et al., 2016	http://regulatorygenomicsgroup.org/chicago RRID:SCR_014941
ChIPQC	Carroll et al., 2019	https://bioconductor.org/packages/release/bioc/html/ChIPQC.html
CRISPOR	Haeussler et al., 2016	http://crispor.tefor.net RRID:SCR_015935
CRISPResso	Pinello et al., 2016	https://github.com/lucapinello/CRISPResso
Cuffdiff	Trapnell et al., 2010	http://cole-trapnell-lab.github.io/cufflinks/cuffdiff/ RRID:SCR_001647
Cufflinks	Trapnell et al., 2010	http://cole-trapnell-lab.github.io/cufflinks/; RRID:SCR_014597
FASTX-Tookit	N/A	http://hannonlab.cshl.edu/fastx_toolkit/; RRID:SCR_005534
Graphpad Prism	GraphPad Software	https://www.graphpad.com/scientific-software/prism/ RRID:SCR_002798
Harmony Software	PerkinElmer	https://www.perkinelmer.com/product/harmony-4-9-office-license-hh17000010
HiC-Pro	Servant et al., 2015	https://github.com/nservant/HiC-Pro RRID:SCR_017643
HOMER	Heinz et al., 2010	http://homer.ucsd.edu/; RRID:SCR_010881
Image Lab	Bio-Rad	https://www.bio-rad.com/en-us/sku/1709690-image-lab-software RRID:SCR_014210
IMPUTE2	Howie et al., 2009	http://mathgen.stats.ox.ac.uk/impute/impute_v2.html; RRID:SCR_013055
LDSC	Bulik-Sullivan et al., 2015; Finucane et al., 2015	https://github.com/bulik/ldsc
MACS2	Zhang et al., 2008	https://github.com/taoliu/MACS/tree/master/MACS2 RRID:SCR_013291
Phantompeaksqualtools	Marinov et al., 2014	https://www.encodeproject.org/software/phantompeakqualtools/ RRID:SCR_005331
SAMtools	Li et al., 2009	http://samtools.sourceforge.net; RRID:SCR_002105
Tophat	Trapnell et al., 2009	http://ccb.jhu.edu/software/tophat/index.shtml; RRID:SCR_013035
Other
7300 Real-Time PCR System	Applied Biosciences	N/A
FACSAria II	BD Biosciences	N/A
Gel Doc XR+	Bio-Rad	N/A
Neon Transfection System	Thermo Fisher Scientific	Cat#MPK5000
Operetta High Content Imaging System	PerkinElmer	N/A
Outside variant resources	This paper	https://github.com/corradin-lab/outside-variants/wiki
QuantStudio 5 Real-Time PCR System	Thermo Fisher Scientific	N/A

Open in a new tab

Highlights.

SNPs that physically interact with a shared target gene often alter MS genetic risk
Outside variant approach predicts pathogenic cell type for individual disease alleles
A subset of MS risk SNPs alter oligodendrocyte-intrinsic cell functions
Dysregulation of RNAPII release confers MS risk and blocks oligodendrocyte maturation

ACKNOWLEDGMENTS

The authors would like to thank Silvi Rouskin and Andrea Cohen for discussion and comments. Additionally, we acknowledge the work of Whitehead Institute’s FACS Core Facility for cell-sorting assistance and Case Western Reserve University’s Genomics Core for RNA and ChIP-seq experiments. This work was supported by NIH grant 5DP1DA044337 for outside-variant methodology development and grants R01 NS096148 and R35 NS097303 for MS tissue work. This investigation was supported grant RG-1701-26707 from the National Multiple Sclerosis Society. This study makes use of data generated by the Blueprint Consortium. A full list of the investigators who contributed to the generation of the data is available at http://www.blueprint-epigenome.eu. Funding for the project was provided by the European Union’s Seventh Framework Programme (FP7/2007-2013) under grant agreement no 282510 BLUEPRINT. This study also makes use of data generated by the Wellcome Trust Case Control Consortium. A full list of the investigators who contributed to the generation of the data is available from https://www.wtccc.org.uk/. Funding for the project was provided by the Wellcome Trust under award 076113

Footnotes

SUPPLEMENTAL INFORMATION

Supplemental Information can be found online at https://doi.org/10.1016/j.cell.2020.03.002.

DECLARATION OF INTERESTS

P.J.T. and D.J.A are co-founders and consultants for Convelo Therapeutics, which has licensed patents unrelated to the current study. P.J.T., D.J.A., and C.W.R.U. retain equity in Convelo Therapeutics. D.C.F. is currently an employee and shareholder of Convelo Therapeutics.

REFERENCES

Adams D, Altucci L, Antonarakis SE, Ballesteros J, Beck S, Bird A, Bock C, Boehm B, Campo E, Caricasole A, et al. (2012). BLUEPRINT to decode the epigenetic signature written in blood. Nat. Biotechnol. 30, 224–226. [DOI] [PubMed] [Google Scholar]
Aschenbrenner D, Foglierini M, Jarrossay D, Hu D, Weiner HL, Kuchroo VK, Lanzavecchia A, Notarbartolo S, and Sallusto F (2018). An immunoregulatory and tissue-residency program modulated by c-MAF in human T_H17 cells. Nat. Immunol. 19, 1126–1136. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al. ; The Gene Ontology Consortium (2000). Gene Ontology: tool for the unification of biology. Nat. Genet. 25, 25–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, Korbel JO, Marchini JL, McCarthy S, McVean GA, and Abecasis GR; 1000 Genomes Project Consortium (2015). A global reference for human genetic variation. Nature 526, 68–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
Baranzini SE, and Oksenberg JR (2017). The genetics of multiple sclerosis: from 0 to 200 in 50 years. Trends Genet. 33, 960–970. [DOI] [PMC free article] [PubMed] [Google Scholar]
Beecham AH, Patsopoulos NA, Xifara DK, Davis MF, Kemppinen A, Cotsapas C, Shah TS, Spencer C, Booth D, Goris A, et al. ; International Multiple Sclerosis Genetics Consortium (IMSGC); Wellcome Trust Case Control Consortium 2 (WTCCC2); International IBD Genetics Consortium (IIBDGC) (2013). Analysis of immune-related loci identifies 48 new susceptibility variants for multiple sclerosis. Nat. Genet. 45, 1353–1360. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bennett JL, and Stijve O (2009). Update on inflammation, neurodegeneration, and immunoregulation in multiple sclerosis: therapeutic implications. Clin. Neuropharmacol. 32, 121–132. [DOI] [PubMed] [Google Scholar]
Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, Duncan L, Perry JR, Patterson N, Robinson EB, et al. ; ReproGen Consortium; Psychiatric Genomics Consortium; Genetic Consortium for Anorexia Nervosa of the Wellcome Trust Case Control Consortium 3 (2015). An atlas of genetic correlations across human diseases and traits. Nat. Genet. 47, 1236–1241. [DOI] [PMC free article] [PubMed] [Google Scholar]
Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, McMahon A, Morales J, Mountjoy E, Sollis E, et al. (2019). The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 47 (D1), D1005–D1012. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cairns J, Freire-Pritchett P, Wingett SW, Varnai C, Dimond A, Plagnol V, Zerbino D, Schoenfelder S, Javierre BM, Osborne C, et al. (2016). CHiCAGO: robust detection of DNA looping interactions in capture Hi-C data. Genome Biol. 17, 127. [DOI] [PMC free article] [PubMed] [Google Scholar]
Canver MC, Lessard S, Pinello L, Wu Y, Ilboudo Y, Stern EN, Needleman AJ, Galacteros F, Brugnara C, Kutlar A, et al. (2017). Variant-aware saturating mutagenesis using multiple Cas9 nucleases identifies regulatory elements at trait-associated loci. Nat. Genet. 49, 625–634. [DOI] [PMC free article] [PubMed] [Google Scholar]
Carroll TS, Liang Z, Salama R, Stark R, and de Santiago I (2014). Impact of artifact removal on ChIP quality metrics in ChIP-seq and ChIP-exo data. Front. Genet. 5, 75. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chang A, Nishiyama A, Peterson J, Prineas J, and Trapp BD (2000). NG2-positive oligodendrocyte progenitor cells in adult human brain and multiple sclerosis lesions. J. Neurosci. 20, 6404–6412. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chang A, Tourtellotte WW, Rudick R, and Trapp BD (2002). Premyelinating oligodendrocytes in chronic lesions of multiple sclerosis. N. Engl. J. Med. 346, 165–173. [DOI] [PubMed] [Google Scholar]
Chen L, Ge B, Casale FP, Vasquez L, Kwan T, Garrido-Martin D, Watt S, Yan Y, Kundu K, Ecker S, et al. (2016). Genetic drivers of epigenetic and transcriptional variation in human immune cells. Cell 167, 1398–1414.e1324. [DOI] [PMC free article] [PubMed] [Google Scholar]
Corradin O, Saiakhova A, Akhtar-Zaidi B, Myeroff L, Willis J, Cowper-Sal lari R, Lupien M, Markowitz S, and Scacheri PC (2014). Combinatorial effects of multiple enhancer variants in linkage disequilibrium dictate levels of gene expression to confer susceptibility to common traits. Genome Res. 24, 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
Corradin O, Cohen AJ, Luppino JM, Bayles IM, Schumacher FR, and Scacheri PC (2016). Modeling disease risk through analysis of physical interactions between genetic variants within chromatin regulatory circuitry. Nat. Genet. 48, 1313–1320. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cotsapas C, Voight BF, Rossin E, Lage K, Neale BM, Wallace C, Abecasis GR, Barrett JC, Behrens T, Cho J, et al. ; FOCiS Network of Consortia (2011). Pervasive sharing of genetic effects in autoimmune disease. PLoS Genet. 7, e1002254. [DOI] [PMC free article] [PubMed] [Google Scholar]
Cusanovich DA, Hill AJ, Aghamirzaie D, Daza RM, Pliner HA, Berletch JB, Filippova GN, Huang X, Christiansen L, DeWitt WS, et al. (2018). A single-cell atlas of in vivo mammalian chromatin accessibility. Cell 174,1309–1324.e1318. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dai J, Zhou S, Ge Q, Qin J, Li J, Ju H, Cao Y, Zheng M, Li C, Gao X, et al. (2019). Recruitment of Brd3 and Brd4to acetylated chromatin is essential for proinflammatory cytokine-induced matrix-degrading enzyme expression. J. Orthop. Surg. Res. 14, 59. [DOI] [PMC free article] [PubMed] [Google Scholar]
De Jager PL, Jia X, Wang J, de Bakker PI, Ottoboni L, Aggarwal NT, Piccio L, Raychaudhuri S, Tran D, Aubin C, et al. ; International MS Genetics Consortium (2009). Meta-analysis of genome scans and replication identify CD6, IRF8 and TNFRSF1A as new multiple sclerosis susceptibility loci. Nat. Genet. 41, 776–782. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dendrou CA, Fugger L, and Friese MA (2015). Immunopathology of multiple sclerosis. Nat. Rev. Immunol. 15, 545–558. [DOI] [PubMed] [Google Scholar]
Deshmukh VA, Tardif V, Lyssiotis CA, Green CC, Kerman B, Kim HJ, Padmanabhan K, Swoboda JG, Ahmad I, Kondo T, et al. (2013). Aregenerative approach to the treatment of multiple sclerosis. Nature 502, 327–332. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dixon JR, Jung I, Selvaraj S, Shen Y, Antosiewicz-Bourget JE, Lee AY, Ye Z, Kim A, Rajagopal N, Xie W, et al. (2015). Chromatin architecture reorganization during stem cell differentiation. Nature 518, 331–336. [DOI] [PMC free article] [PubMed] [Google Scholar]
Dutta R, McDonough J, Yin X, Peterson J, Chang A, Torres T, Gudz T, Macklin WB, Lewis DA, Fox RJ, et al. (2006). Mitochondrial dysfunction as a cause of axonal degeneration in multiple sclerosis patients. Ann. Neurol. 59, 478–489. [DOI] [PubMed] [Google Scholar]
Fabregat A, Jupe S, Matthews L, Sidiropoulos K, Gillespie M, Garapati P, Haw R, Jassal B, Korninger F, May B, et al. (2018). The Reactome pathway knowledgebase. Nucleic Acids Res. 46 (D1), D649–D655. [DOI] [PMC free article] [PubMed] [Google Scholar]
Factor DC, Corradin O, Zentner GE, Saiakhova A, Song L, Chenoweth JG, McKay RD, Crawford GE, Scacheri PC, and Tesar PJ (2014). Epigenomic comparison reveals activation of “seed” enhancers during transition from naive to primed pluripotency. Cell Stem Cell 14, 854–863. [DOI] [PMC free article] [PubMed] [Google Scholar]
Farh KK, Marson A, Zhu J, Kleinewietfeld M, Housley WJ, Beik S, Shoresh N, Whitton H, Ryan RJ, Shishkin AA, et al. (2014). Genetic and epigenetic fine mapping of causal autoimmune disease variants. Nature 518, 337–343. [DOI] [PMC free article] [PubMed] [Google Scholar]
Feinstein A, Freeman J, and Lo AC (2015). Treatment of progressive multiple sclerosis: what works, what does not, and what is needed. Lancet Neurol. 14, 194–207. [DOI] [PubMed] [Google Scholar]
Filippakopoulos P, Qi J, Picaud S, Shen Y, Smith WB, Fedorov O, Morse EM, Keates T, Hickman TT, Felletar I, et al. (2010). Selective inhibition of BET bromodomains. Nature 468, 1067–1073. [DOI] [PMC free article] [PubMed] [Google Scholar]
Filippi M, Bar-Or A, Piehl F, Preziosa P, Solari A, Vukusic S, and Rocca MA (2018). Multiple sclerosis. Nat. Rev. Dis. Primers 4, 43. [DOI] [PubMed] [Google Scholar]
Finucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh PR, Anttila V, Xu H, Zang C, Farh K, et al. ; ReproGen Consortium; Schizophrenia Working Group of the Psychiatric Genomics Consortium; RACI Consortium (2015). Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat. Genet. 47, 1228–1235. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fortini BK, Tring S, Plummer SJ, Edlund CK, Moreno V, Bresalier RS, Barry EL, Church TR, Figueiredo JC, and Casey G (2014). Multiple functional risk variants in a SMAD7 enhancer implicate a colorectal cancer risk haplotype. PLoS ONE 9, e111914. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gandhi R, Laroni A, and Weiner HL (2010). Role of the innate immune system in the pathogenesis of multiple sclerosis. J. Neuroimmunol. 221, 7–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
Glubb DM, Maranian MJ, Michailidou K, Pooley KA, Meyer KB, Kar S, Carlebur S, O’Reilly M, Betts JA, Hillman KM, et al. ; GENICA Network; kConFab Investigators; Norwegian Breast Cancer Study (2015). Fine-scale mapping of the 5q11.2 breast cancer locus reveals at least three independent risk variants regulating MAP3K1. Am. J. Hum. Genet. 96, 5–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
Gosselin D, Skola D, Coufal NG, Holtman IR, Schlachetzki JCM, Sajti E, Jaeger BN, O’Connor C, Fitzpatrick C, Pasillas MP, et al. (2017). An environment-dependent transcriptional network specifies human microglia identity. Science 356, eaa13222. [DOI] [PMC free article] [PubMed] [Google Scholar]
Guo C, Ludvik AE, Arlotto ME, Hayes MG, Armstrong LL, Scholtens DM, Brown CD, Newgard CB, Becker TC, Layden BT, et al. (2015). Coordinated regulatory variation associated with gestational hyperglycaemia regulates expression of the novel hexokinase HKDC1. Nat. Commun. 6, 6069. [DOI] [PMC free article] [PubMed] [Google Scholar]
Haeussler M, Schönig K, Eckert H, Eschstruth A, Mianné J, Renaud JB, Schneider-Maunoury S, Shkumatava A, Teboul L, Kent J, et al. (2016). Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA selection tool CRISPOR. Genome Biol. 17, 148. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hafler DA, Slavik JM, Anderson DE, O’Connor KC, De Jager P, and Baecher-Allan C (2005). Multiple sclerosis. Immunol. Rev. 204, 208–231. [DOI] [PubMed] [Google Scholar]
Hafler DA, Compston A, Sawcer S, Lander ES, Daly MJ, De Jager PL, de Bakker PI, Gabriel SB, Mirel DB, Ivinson AJ, et al. ; International Multiple Sclerosis Genetics Consortium (2007). Risk alleles for multiple sclerosis identified by a genomewide study. N. Engl. J. Med. 357, 851–862. [DOI] [PubMed] [Google Scholar]
Haghikia A, Hohlfeld R, Gold R, and Fugger L (2013). Therapies for multiple sclerosis: translational achievements and outstanding needs. Trends Mol. Med. 19, 309–319. [DOI] [PubMed] [Google Scholar]
He H, Li W, Liyanarachchi S, Srinivas M, Wang Y, Akagi K, Wang Y, Wu D, Wang Q, Jin V, et al. (2015). Multiple functional variants in long-range enhancer elements contribute to the risk of SNP rs965513 in thyroid cancer. Proc. Natl. Acad. Sci. USA 112, 6128–6133. [DOI] [PMC free article] [PubMed] [Google Scholar]
Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, Cheng JX, Murre C, Singh H, and Glass CK (2010). Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hermjakob H, Montecchi-Palazzi L, Lewington C, Mudali S, Kerrien S, Orchard S, Vingron M, Roechert B, Roepstorff P, Valencia A, et al. (2004). IntAct: an open source molecular interaction database. Nucleic Acids Res. 32, D452–D455. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hnisz D, Abraham BJ, Lee TI, Lau A, Saint-André V, Sigova AA, Hoke HA, and Young RA (2013). Super-enhancers in the control of cell identity and disease. Cell 155, 934–947. [DOI] [PMC free article] [PubMed] [Google Scholar]
Horlbeck MA, Gilbert LA, Villalta JE, Adamson B, Pak RA, Chen Y, Fields AP, Park CY, Corn JE, Kampmann M, and Weissman JS (2016). Compact and highly active next-generation libraries for CRISPR-mediated gene repression and activation. eLife 5, e19760. [DOI] [PMC free article] [PubMed] [Google Scholar]
Howie BN, Donnelly P, and Marchini J (2009).A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 5, e1000529. [DOI] [PMC free article] [PubMed] [Google Scholar]
Hubler Z, Allimuthu D, Bederman I, Elitt MS, Madhavan M, Allan KC, Shick HE, Garrison E, T Karl M, Factor DC, et al. (2018). Accumulation of 8,9-unsaturated sterols drives oligodendrocyte formation and remyelination. Nature 560, 372–376. [DOI] [PMC free article] [PubMed] [Google Scholar]
Huynh JL, Garg P, Thin TH, Yoo S, Dutta R, Trapp BD, Haroutunian V, Zhu J, Donovan MJ, Sharp AJ, and Casaccia P (2014). Epigenome-wide differences in pathology-free regions of multiple sclerosis-affected brains. Nat. Neurosci. 17, 121–130. [DOI] [PMC free article] [PubMed] [Google Scholar]
International Multiple Sclerosis GeneticsConsortium (2019). Multiple sclerosis genomic map implicates peripheral immune cells and microglia in susceptibility. Science 365, eaav7188. [DOI] [PMC free article] [PubMed] [Google Scholar]
Javierre BM, Burren OS, Wilder SP, Kreuzhuber R, Hill SM, Sewitz S, Cairns J, Wingett SW, Varnai C, Thiecke MJ, et al. (2016). Lineage-specific genome architecture links enhancers and non-coding disease variants to target gene promoters. Cell 167, 1369–1384.e1319. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jonkers I, and Lis JT (2015). Getting up to speed with transcription elongation by RNA polymerase II. Nat. Rev. Mol. Cell Biol. 16, 167–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
Joung J, Konermann S, Gootenberg JS, Abudayyeh OO, Platt RJ, Brigham MD, Sanjana NE, and Zhang F (2017). Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. Nat. Protoc. 12, 828–863. [DOI] [PMC free article] [PubMed] [Google Scholar]
Jung I, Schmitt A, Diao Y, Lee AJ, Liu T, Yang D, Tan C, Eom J, Chan M, Chee S, et al. (2019). A compendium of promoter-centered long-range chromatin interactions in the human genome. Nat. Genet. 51, 1442–1449. [DOI] [PMC free article] [PubMed] [Google Scholar]
Keirstead HS, and Blakemore WF (1999). The role of oligodendrocytes and oligodendrocyte progenitors in CNS remyelination. Adv. Exp. Med. Biol. 468, 183–197. [DOI] [PubMed] [Google Scholar]
Kozlenkov A, Li J, Apontes P, Hurd YL, Byne WM, Koonin EV, Wegner M, Mukamel EA, and Dracheva S (2018). A unique role for DNA (hydroxy)methylation in epigenetic regulation of human inhibitory neurons. Sci. Adv. 4, eaau6190. [DOI] [PMC free article] [PubMed] [Google Scholar]
Langmead B, and Salzberg SL (2012). Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lengner CJ, Gimelbrant AA, Erwin JA, Cheng AW, Guenther MG, Welstead GG, Alagappan R, Frampton GM, Xu P, Muffat J, et al. (2010). Derivation of pre-X inactivation human embryonic stem cells under physiological oxygen concentrations. Cell 141, 872–883. [DOI] [PubMed] [Google Scholar]
LeRoy G, Rickards B, and Flint SJ (2008). The double bromodomain proteins Brd2 and Brd3 couple histone acetylation to transcription. Mol. Cell 30, 51–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, and Durbin R; 1000 Genome Project Data Processing Subgroup (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079. [DOI] [PMC free article] [PubMed] [Google Scholar]
Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, Powell C, Vedantam S, Buchkovich ML, Yang J, et al. ; LifeLines Cohort Study; ADIPOGen Consortium; AGEN-BMI Working Group; CARDIOGRAMplusC4D Consortium; CKDGen Consortium; GLGC; ICBP; MAGIC Investigators; MuTHER Consortium; MIGen Consortium; PAGE Consortium; ReproGen Consortium; GENIE Consortium; International Endogene Consortium (2015). Genetic studies of body mass indexyield new insights for obesity biology. Nature 518,197–206. [DOI] [PMC free article] [PubMed] [Google Scholar]
Marinov GK, Kundaje A, Park PJ, and Wold BJ (2014). Large-scale quality analysis of published ChIP-seq data. G3 (Bethesda) 4, 209–223. [DOI] [PMC free article] [PubMed] [Google Scholar]
Maurano MT, Humbert R, Rynes E, Thurman RE, Haugen E, Wang H, Reynolds AP, Sandstrom R, Qu H, Brody J, et al. (2012). Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195. [DOI] [PMC free article] [PubMed] [Google Scholar]
McFarland HF, and Martin R (2007). Multiple sclerosis: a complicated picture of autoimmunity. Nat. Immunol. 8, 913–919. [DOI] [PubMed] [Google Scholar]
Mei F, Fancy SPJ, Shen YA, Niu J, Zhao C, Presley B, Miao E, Lee S, Mayoral SR, Redmond SA, et al. (2014). Micropillar arrays as a high-throughput screening platform for therapeutics in multiple sclerosis. Nat. Med. 20, 954–960. [DOI] [PMC free article] [PubMed] [Google Scholar]
Muffat J, Li Y, Yuan B, Mitalipova M, Omer A, Corcoran S, Bakiasi G, Tsai LH, Aubourg P, Ransohoff RM, and Jaenisch R (2016). Efficient derivation of microglia-like cells from human pluripotent stem cells. Nat. Med. 22, 1358–1367. [DOI] [PMC free article] [PubMed] [Google Scholar]
Najm FJ, Zaremba A, Caprariello AV, Nayak S, Freundt EC, Scacheri PC, Miller RH, and Tesar PJ (2011). Rapid and robust generation of functional oligodendrocyte progenitor cells from epiblast stem cells. Nat. Methods 8, 957–962. [DOI] [PMC free article] [PubMed] [Google Scholar]
Najm FJ, Madhavan M, Zaremba A, Shick E, Karl RT, Factor DC, Miller TE, Nevin ZS, Kantor C, Sargent A, et al. (2015). Drug-based modulation of endogenous stem cells promotes functional remyelination in vivo. Nature 522, 216–220. [DOI] [PMC free article] [PubMed] [Google Scholar]
Nevin ZS, Factor DC, Karl RT, Douvaras P, Laukka J, Windrem MS, Goldman SA, Fossati V, Hobson GM, and Tesar PJ (2017). Modeling the mutational and phenotypic landscapes of Pelizaeus-Merzbacher disease with human iPSC-derived oligodendrocytes. Am. J. Hum. Genet. 100, 617–634. [DOI] [PMC free article] [PubMed] [Google Scholar]
Patsopoulos NA, Esposito F, Reischl J, Lehr S, Bauer D, Heubach J, Sandbrink R, Pohl C, Edan G, Kappos L, et al. ; BayerPharma MSGenetics Working Group; Steering Committees of Studies Evaluating IFNβ-1b and a CCR1-Antagonist; ANZgene Consortium; GeneMSA; International Multiple Sclerosis Genetics Consortium (2011). Genome-wide meta-analysis identifies novel multiple sclerosis susceptibility loci. Ann. Neurol. 70, 897–912. [DOI] [PMC free article] [PubMed] [Google Scholar]
Pinello L, Canver MC, Hoban MD, Orkin SH, Kohn DB, Bauer DE, and Yuan GC (2016). Analyzing CRISPR genome-editing experiments with CRISPResso. Nat. Biotechnol. 34, 695–697. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ponath G, Lincoln MR, Levine-Ritterman M, Park C, Dahlawi S, Mubarak M, Sumida T, Airas L, Zhang S, Isitan C, et al. (2018). Enhanced astrocyte responses are driven byagenetic risk allele associated with multiple sclerosis. Nat. Commun. 9, 5337. [DOI] [PMC free article] [PubMed] [Google Scholar]
Rajagopal N, Srinivasan S, Kooshesh K, Guo Y, Edwards MD, Banerjee B, Syed T, Emons BJ, Gifford DK, and Sherwood RI (2016). High-throughput mapping of regulatory DNA. Nat. Biotechnol. 34, 167–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
Ran FA, Hsu PD, Wright J,Agarwala V, Scott DA, and Zhang F (2013). Genome engineering using the CRISPR-Cas9 system. Nat. Protoc. 8, 2281–2308. [DOI] [PMC free article] [PubMed] [Google Scholar]
Sawcer S, Hellenthal G, Pirinen M, Spencer CC, Patsopoulos NA, Moutsianas L, Dilthey A, Su Z, Freeman C, Hunt SE, et al. ; International Multiple Sclerosis Genetics Consortium; Wellcome Trust Case Control Consortium 2 (2011). Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature 476, 214–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schmidt D, Wilson MD, Spyrou C, Brown GD, Hadfield J, and Odom DT (2009). ChIP-seq: using high-throughput sequencing to discover protein-DNA interactions. Methods 48, 240–248. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schmitt AD, Hu M, Jung I, Xu Z, Qiu Y, Tan CL, Li Y, Lin S, Lin Y, Barr CL, and Ren B (2016). A compendium of chromatin contact maps reveals spatially active regions in the human genome. Cell Rep. 17, 2042–2059. [DOI] [PMC free article] [PubMed] [Google Scholar]
Servant N, Varoquaux N, Lajoie BR, Viara E, Chen CJ, Vert JP, Heard E, Dekker J, and Barillot E (2015). HiC-Pro: an optimized and flexible pipeline for Hi-C data processing. Genome Biol. 16, 259. [DOI] [PMC free article] [PubMed] [Google Scholar]
Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, and Pachter L (2010). Transcript assembly and quantification by RNA-seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
Tripathi A, Volsko C, Garcia JP, Agirre E, Allan KC, Tesar PJ, Trapp BD, Castelo-Branco G, Sim FJ, and Dutta R (2019). Oligodendrocyte intrinsic miR-27a controls myelination and remyelination. Cell Rep. 29, 904–919.e909. [DOI] [PMC free article] [PubMed] [Google Scholar]
Yoshida Y, Yoshimi R, Yoshii H, Kim D, Dey A, Xiong H, Munasinghe J, Yazawa I, O’Donovan MJ, Maximova OA, et al. (2014). The transcription factor IRF8 activates integrin-mediated TGF-β signaling and promotes neuroinflammation. Immunity 40, 187–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
Young TM, Tsai M, Tian B, Mathews MB, and Pe’ery T (2007). Cellular mRNA activates transcription elongation by displacing 7SK RNA. PLoS ONE 2, e1010. [DOI] [PMC free article] [PubMed] [Google Scholar]
Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, and Liu XS (2008). Model-based analysis of ChIP-seq (MACS). Genome Biol. 9, R137. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS1581397-supplement-1.pdf^{(789.1KB, pdf)}

NIHMS1581397-supplement-2.pdf^{(1MB, pdf)}

NIHMS1581397-supplement-3.pdf^{(2.1MB, pdf)}

NIHMS1581397-supplement-4.pdf^{(27MB, pdf)}

NIHMS1581397-supplement-5.pdf^{(1.2MB, pdf)}

NIHMS1581397-supplement-6.xlsx^{(49.2KB, xlsx)}

NIHMS1581397-supplement-7.xlsx^{(52.7KB, xlsx)}

NIHMS1581397-supplement-8.pdf^{(2.5MB, pdf)}

Data Availability Statement

https://github.com/corradin-lab/outside-variants

The accession number for the ChIP and RNA-seq data generated for this study is GEO: GSE142084 and GSE142085

[R1] Adams D, Altucci L, Antonarakis SE, Ballesteros J, Beck S, Bird A, Bock C, Boehm B, Campo E, Caricasole A, et al. (2012). BLUEPRINT to decode the epigenetic signature written in blood. Nat. Biotechnol. 30, 224–226. [DOI] [PubMed] [Google Scholar]

[R2] Aschenbrenner D, Foglierini M, Jarrossay D, Hu D, Weiner HL, Kuchroo VK, Lanzavecchia A, Notarbartolo S, and Sallusto F (2018). An immunoregulatory and tissue-residency program modulated by c-MAF in human T_H17 cells. Nat. Immunol. 19, 1126–1136. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R3] Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al. ; The Gene Ontology Consortium (2000). Gene Ontology: tool for the unification of biology. Nat. Genet. 25, 25–29. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R4] Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, Korbel JO, Marchini JL, McCarthy S, McVean GA, and Abecasis GR; 1000 Genomes Project Consortium (2015). A global reference for human genetic variation. Nature 526, 68–74. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R5] Baranzini SE, and Oksenberg JR (2017). The genetics of multiple sclerosis: from 0 to 200 in 50 years. Trends Genet. 33, 960–970. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] Beecham AH, Patsopoulos NA, Xifara DK, Davis MF, Kemppinen A, Cotsapas C, Shah TS, Spencer C, Booth D, Goris A, et al. ; International Multiple Sclerosis Genetics Consortium (IMSGC); Wellcome Trust Case Control Consortium 2 (WTCCC2); International IBD Genetics Consortium (IIBDGC) (2013). Analysis of immune-related loci identifies 48 new susceptibility variants for multiple sclerosis. Nat. Genet. 45, 1353–1360. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R7] Bennett JL, and Stijve O (2009). Update on inflammation, neurodegeneration, and immunoregulation in multiple sclerosis: therapeutic implications. Clin. Neuropharmacol. 32, 121–132. [DOI] [PubMed] [Google Scholar]

[R8] Bulik-Sullivan B, Finucane HK, Anttila V, Gusev A, Day FR, Loh PR, Duncan L, Perry JR, Patterson N, Robinson EB, et al. ; ReproGen Consortium; Psychiatric Genomics Consortium; Genetic Consortium for Anorexia Nervosa of the Wellcome Trust Case Control Consortium 3 (2015). An atlas of genetic correlations across human diseases and traits. Nat. Genet. 47, 1236–1241. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, McMahon A, Morales J, Mountjoy E, Sollis E, et al. (2019). The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 47 (D1), D1005–D1012. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] Cairns J, Freire-Pritchett P, Wingett SW, Varnai C, Dimond A, Plagnol V, Zerbino D, Schoenfelder S, Javierre BM, Osborne C, et al. (2016). CHiCAGO: robust detection of DNA looping interactions in capture Hi-C data. Genome Biol. 17, 127. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] Canver MC, Lessard S, Pinello L, Wu Y, Ilboudo Y, Stern EN, Needleman AJ, Galacteros F, Brugnara C, Kutlar A, et al. (2017). Variant-aware saturating mutagenesis using multiple Cas9 nucleases identifies regulatory elements at trait-associated loci. Nat. Genet. 49, 625–634. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] Carroll TS, Liang Z, Salama R, Stark R, and de Santiago I (2014). Impact of artifact removal on ChIP quality metrics in ChIP-seq and ChIP-exo data. Front. Genet. 5, 75. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] Chang A, Nishiyama A, Peterson J, Prineas J, and Trapp BD (2000). NG2-positive oligodendrocyte progenitor cells in adult human brain and multiple sclerosis lesions. J. Neurosci. 20, 6404–6412. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] Chang A, Tourtellotte WW, Rudick R, and Trapp BD (2002). Premyelinating oligodendrocytes in chronic lesions of multiple sclerosis. N. Engl. J. Med. 346, 165–173. [DOI] [PubMed] [Google Scholar]

[R15] Chen L, Ge B, Casale FP, Vasquez L, Kwan T, Garrido-Martin D, Watt S, Yan Y, Kundu K, Ecker S, et al. (2016). Genetic drivers of epigenetic and transcriptional variation in human immune cells. Cell 167, 1398–1414.e1324. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] Corradin O, Saiakhova A, Akhtar-Zaidi B, Myeroff L, Willis J, Cowper-Sal lari R, Lupien M, Markowitz S, and Scacheri PC (2014). Combinatorial effects of multiple enhancer variants in linkage disequilibrium dictate levels of gene expression to confer susceptibility to common traits. Genome Res. 24, 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] Corradin O, Cohen AJ, Luppino JM, Bayles IM, Schumacher FR, and Scacheri PC (2016). Modeling disease risk through analysis of physical interactions between genetic variants within chromatin regulatory circuitry. Nat. Genet. 48, 1313–1320. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] Cotsapas C, Voight BF, Rossin E, Lage K, Neale BM, Wallace C, Abecasis GR, Barrett JC, Behrens T, Cho J, et al. ; FOCiS Network of Consortia (2011). Pervasive sharing of genetic effects in autoimmune disease. PLoS Genet. 7, e1002254. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] Cusanovich DA, Hill AJ, Aghamirzaie D, Daza RM, Pliner HA, Berletch JB, Filippova GN, Huang X, Christiansen L, DeWitt WS, et al. (2018). A single-cell atlas of in vivo mammalian chromatin accessibility. Cell 174,1309–1324.e1318. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] Dai J, Zhou S, Ge Q, Qin J, Li J, Ju H, Cao Y, Zheng M, Li C, Gao X, et al. (2019). Recruitment of Brd3 and Brd4to acetylated chromatin is essential for proinflammatory cytokine-induced matrix-degrading enzyme expression. J. Orthop. Surg. Res. 14, 59. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] De Jager PL, Jia X, Wang J, de Bakker PI, Ottoboni L, Aggarwal NT, Piccio L, Raychaudhuri S, Tran D, Aubin C, et al. ; International MS Genetics Consortium (2009). Meta-analysis of genome scans and replication identify CD6, IRF8 and TNFRSF1A as new multiple sclerosis susceptibility loci. Nat. Genet. 41, 776–782. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R22] Dendrou CA, Fugger L, and Friese MA (2015). Immunopathology of multiple sclerosis. Nat. Rev. Immunol. 15, 545–558. [DOI] [PubMed] [Google Scholar]

[R23] Deshmukh VA, Tardif V, Lyssiotis CA, Green CC, Kerman B, Kim HJ, Padmanabhan K, Swoboda JG, Ahmad I, Kondo T, et al. (2013). Aregenerative approach to the treatment of multiple sclerosis. Nature 502, 327–332. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R24] Dixon JR, Jung I, Selvaraj S, Shen Y, Antosiewicz-Bourget JE, Lee AY, Ye Z, Kim A, Rajagopal N, Xie W, et al. (2015). Chromatin architecture reorganization during stem cell differentiation. Nature 518, 331–336. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] Dutta R, McDonough J, Yin X, Peterson J, Chang A, Torres T, Gudz T, Macklin WB, Lewis DA, Fox RJ, et al. (2006). Mitochondrial dysfunction as a cause of axonal degeneration in multiple sclerosis patients. Ann. Neurol. 59, 478–489. [DOI] [PubMed] [Google Scholar]

[R26] Fabregat A, Jupe S, Matthews L, Sidiropoulos K, Gillespie M, Garapati P, Haw R, Jassal B, Korninger F, May B, et al. (2018). The Reactome pathway knowledgebase. Nucleic Acids Res. 46 (D1), D649–D655. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R27] Factor DC, Corradin O, Zentner GE, Saiakhova A, Song L, Chenoweth JG, McKay RD, Crawford GE, Scacheri PC, and Tesar PJ (2014). Epigenomic comparison reveals activation of “seed” enhancers during transition from naive to primed pluripotency. Cell Stem Cell 14, 854–863. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] Farh KK, Marson A, Zhu J, Kleinewietfeld M, Housley WJ, Beik S, Shoresh N, Whitton H, Ryan RJ, Shishkin AA, et al. (2014). Genetic and epigenetic fine mapping of causal autoimmune disease variants. Nature 518, 337–343. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] Feinstein A, Freeman J, and Lo AC (2015). Treatment of progressive multiple sclerosis: what works, what does not, and what is needed. Lancet Neurol. 14, 194–207. [DOI] [PubMed] [Google Scholar]

[R30] Filippakopoulos P, Qi J, Picaud S, Shen Y, Smith WB, Fedorov O, Morse EM, Keates T, Hickman TT, Felletar I, et al. (2010). Selective inhibition of BET bromodomains. Nature 468, 1067–1073. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] Filippi M, Bar-Or A, Piehl F, Preziosa P, Solari A, Vukusic S, and Rocca MA (2018). Multiple sclerosis. Nat. Rev. Dis. Primers 4, 43. [DOI] [PubMed] [Google Scholar]

[R32] Finucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh PR, Anttila V, Xu H, Zang C, Farh K, et al. ; ReproGen Consortium; Schizophrenia Working Group of the Psychiatric Genomics Consortium; RACI Consortium (2015). Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat. Genet. 47, 1228–1235. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R33] Fortini BK, Tring S, Plummer SJ, Edlund CK, Moreno V, Bresalier RS, Barry EL, Church TR, Figueiredo JC, and Casey G (2014). Multiple functional risk variants in a SMAD7 enhancer implicate a colorectal cancer risk haplotype. PLoS ONE 9, e111914. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R34] Gandhi R, Laroni A, and Weiner HL (2010). Role of the innate immune system in the pathogenesis of multiple sclerosis. J. Neuroimmunol. 221, 7–14. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R35] Glubb DM, Maranian MJ, Michailidou K, Pooley KA, Meyer KB, Kar S, Carlebur S, O’Reilly M, Betts JA, Hillman KM, et al. ; GENICA Network; kConFab Investigators; Norwegian Breast Cancer Study (2015). Fine-scale mapping of the 5q11.2 breast cancer locus reveals at least three independent risk variants regulating MAP3K1. Am. J. Hum. Genet. 96, 5–20. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] Gosselin D, Skola D, Coufal NG, Holtman IR, Schlachetzki JCM, Sajti E, Jaeger BN, O’Connor C, Fitzpatrick C, Pasillas MP, et al. (2017). An environment-dependent transcriptional network specifies human microglia identity. Science 356, eaa13222. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] Guo C, Ludvik AE, Arlotto ME, Hayes MG, Armstrong LL, Scholtens DM, Brown CD, Newgard CB, Becker TC, Layden BT, et al. (2015). Coordinated regulatory variation associated with gestational hyperglycaemia regulates expression of the novel hexokinase HKDC1. Nat. Commun. 6, 6069. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] Haeussler M, Schönig K, Eckert H, Eschstruth A, Mianné J, Renaud JB, Schneider-Maunoury S, Shkumatava A, Teboul L, Kent J, et al. (2016). Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA selection tool CRISPOR. Genome Biol. 17, 148. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] Hafler DA, Slavik JM, Anderson DE, O’Connor KC, De Jager P, and Baecher-Allan C (2005). Multiple sclerosis. Immunol. Rev. 204, 208–231. [DOI] [PubMed] [Google Scholar]

[R40] Hafler DA, Compston A, Sawcer S, Lander ES, Daly MJ, De Jager PL, de Bakker PI, Gabriel SB, Mirel DB, Ivinson AJ, et al. ; International Multiple Sclerosis Genetics Consortium (2007). Risk alleles for multiple sclerosis identified by a genomewide study. N. Engl. J. Med. 357, 851–862. [DOI] [PubMed] [Google Scholar]

[R41] Haghikia A, Hohlfeld R, Gold R, and Fugger L (2013). Therapies for multiple sclerosis: translational achievements and outstanding needs. Trends Mol. Med. 19, 309–319. [DOI] [PubMed] [Google Scholar]

[R42] He H, Li W, Liyanarachchi S, Srinivas M, Wang Y, Akagi K, Wang Y, Wu D, Wang Q, Jin V, et al. (2015). Multiple functional variants in long-range enhancer elements contribute to the risk of SNP rs965513 in thyroid cancer. Proc. Natl. Acad. Sci. USA 112, 6128–6133. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, Cheng JX, Murre C, Singh H, and Glass CK (2010). Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] Hermjakob H, Montecchi-Palazzi L, Lewington C, Mudali S, Kerrien S, Orchard S, Vingron M, Roechert B, Roepstorff P, Valencia A, et al. (2004). IntAct: an open source molecular interaction database. Nucleic Acids Res. 32, D452–D455. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] Hnisz D, Abraham BJ, Lee TI, Lau A, Saint-André V, Sigova AA, Hoke HA, and Young RA (2013). Super-enhancers in the control of cell identity and disease. Cell 155, 934–947. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] Horlbeck MA, Gilbert LA, Villalta JE, Adamson B, Pak RA, Chen Y, Fields AP, Park CY, Corn JE, Kampmann M, and Weissman JS (2016). Compact and highly active next-generation libraries for CRISPR-mediated gene repression and activation. eLife 5, e19760. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] Howie BN, Donnelly P, and Marchini J (2009).A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 5, e1000529. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] Hubler Z, Allimuthu D, Bederman I, Elitt MS, Madhavan M, Allan KC, Shick HE, Garrison E, T Karl M, Factor DC, et al. (2018). Accumulation of 8,9-unsaturated sterols drives oligodendrocyte formation and remyelination. Nature 560, 372–376. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] Huynh JL, Garg P, Thin TH, Yoo S, Dutta R, Trapp BD, Haroutunian V, Zhu J, Donovan MJ, Sharp AJ, and Casaccia P (2014). Epigenome-wide differences in pathology-free regions of multiple sclerosis-affected brains. Nat. Neurosci. 17, 121–130. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R50] International Multiple Sclerosis GeneticsConsortium (2019). Multiple sclerosis genomic map implicates peripheral immune cells and microglia in susceptibility. Science 365, eaav7188. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] Javierre BM, Burren OS, Wilder SP, Kreuzhuber R, Hill SM, Sewitz S, Cairns J, Wingett SW, Varnai C, Thiecke MJ, et al. (2016). Lineage-specific genome architecture links enhancers and non-coding disease variants to target gene promoters. Cell 167, 1369–1384.e1319. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] Jonkers I, and Lis JT (2015). Getting up to speed with transcription elongation by RNA polymerase II. Nat. Rev. Mol. Cell Biol. 16, 167–177. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R53] Joung J, Konermann S, Gootenberg JS, Abudayyeh OO, Platt RJ, Brigham MD, Sanjana NE, and Zhang F (2017). Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening. Nat. Protoc. 12, 828–863. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] Jung I, Schmitt A, Diao Y, Lee AJ, Liu T, Yang D, Tan C, Eom J, Chan M, Chee S, et al. (2019). A compendium of promoter-centered long-range chromatin interactions in the human genome. Nat. Genet. 51, 1442–1449. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] Keirstead HS, and Blakemore WF (1999). The role of oligodendrocytes and oligodendrocyte progenitors in CNS remyelination. Adv. Exp. Med. Biol. 468, 183–197. [DOI] [PubMed] [Google Scholar]

[R56] Kozlenkov A, Li J, Apontes P, Hurd YL, Byne WM, Koonin EV, Wegner M, Mukamel EA, and Dracheva S (2018). A unique role for DNA (hydroxy)methylation in epigenetic regulation of human inhibitory neurons. Sci. Adv. 4, eaau6190. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R57] Langmead B, and Salzberg SL (2012). Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R58] Lengner CJ, Gimelbrant AA, Erwin JA, Cheng AW, Guenther MG, Welstead GG, Alagappan R, Frampton GM, Xu P, Muffat J, et al. (2010). Derivation of pre-X inactivation human embryonic stem cells under physiological oxygen concentrations. Cell 141, 872–883. [DOI] [PubMed] [Google Scholar]

[R59] LeRoy G, Rickards B, and Flint SJ (2008). The double bromodomain proteins Brd2 and Brd3 couple histone acetylation to transcription. Mol. Cell 30, 51–60. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R60] Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, and Durbin R; 1000 Genome Project Data Processing Subgroup (2009). The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R61] Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, Powell C, Vedantam S, Buchkovich ML, Yang J, et al. ; LifeLines Cohort Study; ADIPOGen Consortium; AGEN-BMI Working Group; CARDIOGRAMplusC4D Consortium; CKDGen Consortium; GLGC; ICBP; MAGIC Investigators; MuTHER Consortium; MIGen Consortium; PAGE Consortium; ReproGen Consortium; GENIE Consortium; International Endogene Consortium (2015). Genetic studies of body mass indexyield new insights for obesity biology. Nature 518,197–206. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R62] Marinov GK, Kundaje A, Park PJ, and Wold BJ (2014). Large-scale quality analysis of published ChIP-seq data. G3 (Bethesda) 4, 209–223. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R63] Maurano MT, Humbert R, Rynes E, Thurman RE, Haugen E, Wang H, Reynolds AP, Sandstrom R, Qu H, Brody J, et al. (2012). Systematic localization of common disease-associated variation in regulatory DNA. Science 337, 1190–1195. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R64] McFarland HF, and Martin R (2007). Multiple sclerosis: a complicated picture of autoimmunity. Nat. Immunol. 8, 913–919. [DOI] [PubMed] [Google Scholar]

[R65] Mei F, Fancy SPJ, Shen YA, Niu J, Zhao C, Presley B, Miao E, Lee S, Mayoral SR, Redmond SA, et al. (2014). Micropillar arrays as a high-throughput screening platform for therapeutics in multiple sclerosis. Nat. Med. 20, 954–960. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R66] Muffat J, Li Y, Yuan B, Mitalipova M, Omer A, Corcoran S, Bakiasi G, Tsai LH, Aubourg P, Ransohoff RM, and Jaenisch R (2016). Efficient derivation of microglia-like cells from human pluripotent stem cells. Nat. Med. 22, 1358–1367. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R67] Najm FJ, Zaremba A, Caprariello AV, Nayak S, Freundt EC, Scacheri PC, Miller RH, and Tesar PJ (2011). Rapid and robust generation of functional oligodendrocyte progenitor cells from epiblast stem cells. Nat. Methods 8, 957–962. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R68] Najm FJ, Madhavan M, Zaremba A, Shick E, Karl RT, Factor DC, Miller TE, Nevin ZS, Kantor C, Sargent A, et al. (2015). Drug-based modulation of endogenous stem cells promotes functional remyelination in vivo. Nature 522, 216–220. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R69] Nevin ZS, Factor DC, Karl RT, Douvaras P, Laukka J, Windrem MS, Goldman SA, Fossati V, Hobson GM, and Tesar PJ (2017). Modeling the mutational and phenotypic landscapes of Pelizaeus-Merzbacher disease with human iPSC-derived oligodendrocytes. Am. J. Hum. Genet. 100, 617–634. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R70] Patsopoulos NA, Esposito F, Reischl J, Lehr S, Bauer D, Heubach J, Sandbrink R, Pohl C, Edan G, Kappos L, et al. ; BayerPharma MSGenetics Working Group; Steering Committees of Studies Evaluating IFNβ-1b and a CCR1-Antagonist; ANZgene Consortium; GeneMSA; International Multiple Sclerosis Genetics Consortium (2011). Genome-wide meta-analysis identifies novel multiple sclerosis susceptibility loci. Ann. Neurol. 70, 897–912. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R71] Pinello L, Canver MC, Hoban MD, Orkin SH, Kohn DB, Bauer DE, and Yuan GC (2016). Analyzing CRISPR genome-editing experiments with CRISPResso. Nat. Biotechnol. 34, 695–697. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R72] Ponath G, Lincoln MR, Levine-Ritterman M, Park C, Dahlawi S, Mubarak M, Sumida T, Airas L, Zhang S, Isitan C, et al. (2018). Enhanced astrocyte responses are driven byagenetic risk allele associated with multiple sclerosis. Nat. Commun. 9, 5337. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R73] Rajagopal N, Srinivasan S, Kooshesh K, Guo Y, Edwards MD, Banerjee B, Syed T, Emons BJ, Gifford DK, and Sherwood RI (2016). High-throughput mapping of regulatory DNA. Nat. Biotechnol. 34, 167–174. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R74] Ran FA, Hsu PD, Wright J,Agarwala V, Scott DA, and Zhang F (2013). Genome engineering using the CRISPR-Cas9 system. Nat. Protoc. 8, 2281–2308. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R75] Sawcer S, Hellenthal G, Pirinen M, Spencer CC, Patsopoulos NA, Moutsianas L, Dilthey A, Su Z, Freeman C, Hunt SE, et al. ; International Multiple Sclerosis Genetics Consortium; Wellcome Trust Case Control Consortium 2 (2011). Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature 476, 214–219. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R76] Schmidt D, Wilson MD, Spyrou C, Brown GD, Hadfield J, and Odom DT (2009). ChIP-seq: using high-throughput sequencing to discover protein-DNA interactions. Methods 48, 240–248. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R77] Schmitt AD, Hu M, Jung I, Xu Z, Qiu Y, Tan CL, Li Y, Lin S, Lin Y, Barr CL, and Ren B (2016). A compendium of chromatin contact maps reveals spatially active regions in the human genome. Cell Rep. 17, 2042–2059. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R78] Servant N, Varoquaux N, Lajoie BR, Viara E, Chen CJ, Vert JP, Heard E, Dekker J, and Barillot E (2015). HiC-Pro: an optimized and flexible pipeline for Hi-C data processing. Genome Biol. 16, 259. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R79] Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, and Pachter L (2010). Transcript assembly and quantification by RNA-seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R80] Tripathi A, Volsko C, Garcia JP, Agirre E, Allan KC, Tesar PJ, Trapp BD, Castelo-Branco G, Sim FJ, and Dutta R (2019). Oligodendrocyte intrinsic miR-27a controls myelination and remyelination. Cell Rep. 29, 904–919.e909. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R81] Yoshida Y, Yoshimi R, Yoshii H, Kim D, Dey A, Xiong H, Munasinghe J, Yazawa I, O’Donovan MJ, Maximova OA, et al. (2014). The transcription factor IRF8 activates integrin-mediated TGF-β signaling and promotes neuroinflammation. Immunity 40, 187–198. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R82] Young TM, Tsai M, Tian B, Mathews MB, and Pe’ery T (2007). Cellular mRNA activates transcription elongation by displacing 7SK RNA. PLoS ONE 2, e1010. [DOI] [PMC free article] [PubMed] [Google Scholar]

[R83] Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, and Liu XS (2008). Model-based analysis of ChIP-seq (MACS). Genome Biol. 9, R137. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Cell-Type-Specific Intralocus Interactions Reveal Oligodendrocyte Mechanisms in MS

Daniel C Factor

Anna M Barbeau

Kevin C Allan

Lucille R Hu

Mayur Madhavan

An T Hoang

Kathryn EA Hazel

Parker A Hall

Sagar Nisraiyya

Fadi J Najm

Tyler E Miller

Zachary S Nevin

Robert T Karl

Bruna R Lima

Yanwei Song

Alexandra G Sibert

Gursimran K Dhillon

Christina Volsko

Cynthia F Bartels

Drew J Adams

Ranjan Dutta

Michael D Gallagher

William Phu

Alexey Kozlenkov

Stella Dracheva

Peter C Scacheri

Paul J Tesar

Olivia Corradin

SUMMARY

Graphical Abstract

In Brief

INTRODUCTION

Figure 1. Cell-Type Specificity of Intralocus Regulatory Elements Enables Prediction of Pathogenic Cell Type of MS Risk Loci.

RESULTS

Cell-Type Specificity of Intralocus Regulatory Elements Enables Prediction of Pathogenic Cell Type of MS Risk Loci

Oligodendrocyte-Specific Regulation of Transcriptional Pause Release Factors Are Associated with MS Risk

Figure 2. Oligodendrocyte-Specific Regulation of Transcriptional Pause Release Factors Is Associated with MS Risk.

Inhibition of Transcriptional Pause Release Blocks Oligodendrocyte Maturation

Figure 3. Inhibition of Transcriptional Pause Release Blocks Oligodendrocyte Maturation.

Dysregulation of Hexim1 Alters Oligodendrocyte Maturation

Figure 4. Dysregulation of Hexim1 Alters Oligodendrocyte Maturation.

Transcriptional Elongation Factors Are Dysregulated in MS Patient Brains

Figure 5. Transcriptional Elongation Factors Are Dysregulated in MS Patient White Matter.

DISCUSSION

STAR★METHODS

LEAD CONTACT AND MATERIALS AVAILABILITY

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines

Jurkat cells

GM12878 cells

Primary cell culture

Generation of neural progenitor cells

Mouse OPC culture

Primary mouse OPC culture

hOPC culture

DRG co-culture

Multiple sclerosis patient tissue

Multiple sclerosis genotyped samples

METHOD DETAILS

Hi-C data analysis

Curation of ChIP-seq panel

Data acquisition, processing and QC

Consensus peak list

Outside variant approach

Definition of candidate outside variants

Identifying outside variants that alter risk

Adaptive permutations

Multi-test correction

Pathogenic cell type prediction

Overview

Quantifying ChIP-seq differences

LD score regression

Overview

Annotated SNP lists

H3K27ac global enrichment

Outside variant defined clusters

Transcriptional pausing pathway

CRISPR/Cas9 outside variants experiments