FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2604964915> ?p ?o ?g. }

• W2604964915 endingPage "570.e10" @default.
• W2604964915 startingPage "558" @default.
• W2604964915 abstract "•Genome-wide maps of eQTL/ASE genes in a multi-ethnic cohort of 91 healthy donors•Reporters screen candidate functional variants in blood lipid-associated eQTLs•CRISPR-Cas9 in hPSCs and mice define functional variants of lipid-associated eQTLs•Mouse models establish HLC eQTL/ASE genes as lipid-functional genes Genome-wide association studies have struggled to identify functional genes and variants underlying complex phenotypes. We recruited a multi-ethnic cohort of healthy volunteers (n = 91) and used their tissue to generate induced pluripotent stem cells (iPSCs) and hepatocyte-like cells (HLCs) for genome-wide mapping of expression quantitative trait loci (eQTLs) and allele-specific expression (ASE). We identified many eQTL genes (eGenes) not observed in the comparably sized Genotype-Tissue Expression project’s human liver cohort (n = 96). Focusing on blood lipid-associated loci, we performed massively parallel reporter assays to screen candidate functional variants and used genome-edited stem cells, CRISPR interference, and mouse modeling to establish rs2277862-CPNE1, rs10889356-DOCK7, rs10889356-ANGPTL3, and rs10872142-FRK as functional SNP-gene sets. We demonstrated HLC eGenes CPNE1, VKORC1, UBE2L3, and ANGPTL3 and HLC ASE gene ACAA2 to be lipid-functional genes in mouse models. These findings endorse an iPSC-based experimental framework to discover functional variants and genes contributing to complex human traits. Genome-wide association studies have struggled to identify functional genes and variants underlying complex phenotypes. We recruited a multi-ethnic cohort of healthy volunteers (n = 91) and used their tissue to generate induced pluripotent stem cells (iPSCs) and hepatocyte-like cells (HLCs) for genome-wide mapping of expression quantitative trait loci (eQTLs) and allele-specific expression (ASE). We identified many eQTL genes (eGenes) not observed in the comparably sized Genotype-Tissue Expression project’s human liver cohort (n = 96). Focusing on blood lipid-associated loci, we performed massively parallel reporter assays to screen candidate functional variants and used genome-edited stem cells, CRISPR interference, and mouse modeling to establish rs2277862-CPNE1, rs10889356-DOCK7, rs10889356-ANGPTL3, and rs10872142-FRK as functional SNP-gene sets. We demonstrated HLC eGenes CPNE1, VKORC1, UBE2L3, and ANGPTL3 and HLC ASE gene ACAA2 to be lipid-functional genes in mouse models. These findings endorse an iPSC-based experimental framework to discover functional variants and genes contributing to complex human traits. Genome-wide association studies (GWASs) have emerged as a robust unbiased approach to identify SNPs associated with incidence of a particular phenotype or disease (Manolio, 2010Manolio T.A. Genomewide association studies and assessment of the risk of disease.N. Engl. J. Med. 2010; 363: 166-176Crossref PubMed Scopus (1116) Google Scholar). Only a small fraction of GWAS lead variants lie within coding sequence and thus directly implicate a functional gene at a locus; the vast majority of lead SNPs fall in noncoding sequence. Moreover, most of these SNPs are not themselves functional but exist in linkage disequilibrium (LD) with the true functional variants. Because many human disease-associated variants are believed to regulate gene expression, expression quantitative trait locus (eQTL) and allele-specific expression (ASE) studies may illuminate potential downstream targets of functional variants. These regulated genes then become candidates for experimental manipulation to ascertain their relevance to the phenotype of interest. However, functional studies elucidating the mechanisms of identified variants have remained a challenge due to the need for laborious experiments and the lack of suitable model systems for noncoding sequence studies. Recently emergent technologies make it feasible to identify and interrogate the function of noncoding variants at eQTL and ASE loci in human model systems. Human pluripotent stem cells (hPSCs), especially induced pluripotent stem cells (iPSCs), make it possible to generate cohorts of person-specific, renewable, differentiated cell lines in vitro (Zhu et al., 2011Zhu H. Lensch M.W. Cahan P. Daley G.Q. Investigating monogenic and complex diseases with pluripotent stem cells.Nat. Rev. Genet. 2011; 12: 266-275Crossref PubMed Scopus (90) Google Scholar). In theory, when drawn from a population with diverse genotypes of common genetic variants, these cohorts might offer the opportunity to validate known eQTL/ASE loci and discover new eQTL/ASE loci “in the dish.” Massively parallel reporter assays (MPRAs) allow investigators to generate high-complexity pools of reporter constructs where each regulatory element or variant of interest is linked to a synthetic reporter gene that carries an identifying barcode (Melnikov et al., 2012Melnikov A. Murugan A. Zhang X. Tesileanu T. Wang L. Rogov P. Feizi S. Gnirke A. Callan Jr., C.G. Kinney J.B. et al.Systematic dissection and optimization of inducible enhancers in human cells using a massively parallel reporter assay.Nat. Biotechnol. 2012; 30: 271-277Crossref PubMed Scopus (389) Google Scholar, Patwardhan et al., 2012Patwardhan R.P. Hiatt J.B. Witten D.M. Kim M.J. Smith R.P. May D. Lee C. Andrie J.M. Lee S.I. Cooper G.M. et al.Massively parallel functional dissection of mammalian enhancers in vivo.Nat. Biotechnol. 2012; 30: 265-270Crossref PubMed Scopus (325) Google Scholar). The reporter construct pools are introduced into cells, and the relative transcriptional activities of the individual elements or variants are measured by sequencing the transcribed reporter mRNAs and counting their specific barcodes. This approach can be used to rapidly profile the regulatory activity of thousands of variants at GWAS loci (Tewhey et al., 2016Tewhey R. Kotliar D. Park D.S. Liu B. Winnicki S. Reilly S.K. Andersen K.G. Mikkelsen T.S. Lander E.S. Schaffner S.F. Sabeti P.C. Direct identification of hundreds of expression-modulating variants using a multiplexed reporter assay.Cell. 2016; 165: 1519-1529Abstract Full Text Full Text PDF PubMed Scopus (205) Google Scholar, Ulirsch et al., 2016Ulirsch J.C. Nandakumar S.K. Wang L. Giani F.C. Zhang X. Rogov P. Melnikov A. McDonel P. Do R. Mikkelsen T.S. Sankaran V.G. Systematic functional dissection of common genetic variation affecting red blood cell traits.Cell. 2016; 165: 1530-1545Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar). Finally, advances in genome-editing technologies—most notably clustered regularly interspaced short palindromic repeats (CRISPR)-CRISPR-associated 9 (Cas9) systems—have opened up new avenues to rigorously assess the functional impact of genetic variation (Musunuru, 2013Musunuru K. Genome editing of human pluripotent stem cells to generate human cellular disease models.Dis. Model. Mech. 2013; 6: 896-904Crossref PubMed Scopus (115) Google Scholar). In this study, we asked two overarching questions. First, can population cohorts of iPSCs and iPSC-differentiated cells be used to perform unbiased genome-wide eQTL/ASE studies in a manner that is complementary to traditional primary tissue-based studies such as the Gene-Tissue Expression (GTEx) project? Second, can we better understand the functional role of human genetic variation in influencing quantitative phenotypic traits, particularly those related to liver metabolism such as blood lipid levels? As part of the NHLBI Next Generation Genetic Association Studies Consortium, we generated population-based cohorts of iPSCs and iPSC-differentiated hepatocyte-like cells (HLCs) to perform genome-wide mapping and characterize known and new eQTL/ASE loci. We thereafter employed gene overexpression mouse models as well as a combination of MPRAs and CRISPR-Cas9 in hPSCs, other types of cultured cells, and mouse models to screen, identify, and validate functional variants and/or genes in several blood lipid-associated eQTL/ASE loci. We generated iPSCs from peripheral blood mononuclear cells isolated from 91 individuals, predominantly African Americans (43%) and European Americans (55%), with more women (60%) than men (40%) (Table S1). All established iPSC lines were confirmed to be free of exogenous Sendai viral reprogramming factor expression and then tested for pluripotency by fluorescence-activated cell sorting (FACS) staining for SSEA4 and Tra-1-60 (Figure S1A; Table S1). Samples passing these criteria were differentiated into HLCs. Differentiated HLCs were morphologically similar to primary hepatocytes, and for a subset of HLC samples, expression of HNF4α was confirmed by immunofluorescence (Figure S1B). The HLCs also secreted apolipoprotein B (apoB), triglycerides, and albumin (Figure S2A), a set of properties unique to functional hepatocytes. In addition to molecular and qualitative assessments of pluripotency and differentiation efficiencies, we used computational means to further validate the iPSCs and HLCs used for our study. We generated RNA sequencing (RNA-seq) data from iPSCs (n = 89), HLCs (n = 86), and primary human hepatocytes (n = 4). In addition, we used RNA-seq data of human whole-liver samples from GTEx (n = 96) (Aguet et al., 2016Aguet F. Brown A.A. Castel S. Davis J.R. Mohammadi P. Segre A.V. Zappala Z. Abell N.S. Fresard L. Gamazon E.R. et al.Local genetic effects on gene expression across 44 human tissues.bioRxiv. 2016; https://doi.org/10.1101/074450Crossref Google Scholar). With the use of singular value decomposition (SVD), expression levels of a panel of differentiation markers (Carcamo-Orive et al., 2017Carcamo-Orive I. Hoffman G.E. Cundiff P. Beckmann N.D. D’Souza S.L. Knowles J.W. Patel A. Papatsenko D. Abbasi F. Reaven G.M. et al.Analysis of transcriptional variability in a large human iPSC library reveals genetic and non-genetic determinants of heterogeneity.Cell Stem Cell. 2017; 20 (this issue): 518-532Abstract Full Text Full Text PDF PubMed Scopus (158) Google Scholar) demonstrated that our iPSCs were extremely similar to other iPSCs and human embryonic stem cells (ESCs) with previously published gene expression data (Choi et al., 2015Choi J. Lee S. Mallard W. Clement K. Tagliazucchi G.M. Lim H. Choi I.Y. Ferrari F. Tsankov A.M. Pop R. et al.A comparison of genetically matched cell lines reveals the equivalence of human iPSCs and ESCs.Nat. Biotechnol. 2015; 33: 1173-1181Crossref PubMed Scopus (181) Google Scholar) (Figure S3A). We characterized the global gene expression profiles of the iPSCs, HLCs, primary hepatocytes, and GTEx livers and in general found that the HLCs had been successfully differentiated from the iPSCs and approximated the gene expression profiles of the primary hepatocytes more closely than the profiles of the livers (Figure S3B). We specifically assessed the expression levels of six hepatocyte-specific genes in the HLCs (Figure S2B), using these levels as compared to levels in primary human hepatocytes as a means to assess the quality of the HLCs (scored from zero to five) (Table S1). Fully understanding the underlying variance between HLCs and whole livers merits further investigation, as hepatocytes constitute only one of several cell types in the liver. We quantified differential gene expression using DESeq2 (Love et al., 2014Love M.I. Huber W. Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2.Genome Biol. 2014; 15: 550Crossref PubMed Scopus (32370) Google Scholar). We identified 3,709 genes more highly expressed in livers compared to iPSCs and 2,040 genes more highly expressed in HLCs compared to iPSCs at false discovery rate (FDR) < 5%. We sought to assess the liver-specific functions associated with 1,349 genes that were common to both sets (differentially expressed in both HLCs and livers compared to iPSCs) by estimating the enrichment of these genes for gene ontology (GO) terms. We identified 126 common GO terms including functions specific to hepatic regulation and lipid metabolism (Table S2). With a goal of identifying hepatic cis-regulatory variation, we mapped genome-wide cis-eQTLs in the iPSCs, HLCs, and GTEx livers using linear models as implemented in Matrix eQTL (Shabalin, 2012Shabalin A.A. Matrix eQTL: ultra fast eQTL analysis via large matrix operations.Bioinformatics. 2012; 28: 1353-1358Crossref PubMed Scopus (837) Google Scholar) (Figure 1A). We identified a total of 3,587 eGenes in the iPSCs, 2,392 eGenes in the HLCs, and 1,511 eGenes in the GTEx livers at FDR < 5% (Table S3). The HLCs and livers shared a total of 477 eGenes, 176 of which were not associated in the iPSCs at FDR < 5% (Figure 1B). The eQTLs were primarily associated with protein coding genes (83%), and the majority of eGenes were nominally associated (p < 0.05) in more than one cell type. To complement the eQTL analyses, and to assess the effects of cis-regulatory variants at lower frequencies that eQTL analyses are underpowered to detect, we quantified allele-specific expression (ASE) in iPSCs and HLCs (individual-level data not provided in the manuscript due to privacy concerns). Genome-wide, we identified 1,721 independent SNPs in 494 genes that exhibited significant ASE in iPSCs (FDR < 5%) and 2,137 independent SNPs in 631 genes that exhibited significant ASE in HLCs. Furthermore, 876 independent SNPs in 541 genes exhibited differential ASE (FDR < 5%) between iPSCs and HLCs (Table S4), indicating cell-type-specific regulatory effects. Genes exhibiting differential ASE were significantly enriched for cell-type-specific eGenes (odds ratio = 3.3, p < 1.8 × 10−13), independently confirming the cell-type-specific regulatory effects identified with the two methodologies. To further refine cell-type-specific effects of cis-eQTLs between iPSCs and HLCs, we conducted a mixed model meta-analysis. Considering our paired sample design, we used METASOFT (Han and Eskin, 2012Han B. Eskin E. Interpreting meta-analyses of genome-wide association studies.PLoS Genet. 2012; 8: e1002555Crossref PubMed Scopus (104) Google Scholar) and Meta-Tissue (Sul et al., 2013Sul J.H. Han B. Ye C. Choi T. Eskin E. Effectively identifying eQTLs from multiple tissues by combining mixed model and meta-analytic approaches.PLoS Genet. 2013; 9: e1003491Crossref PubMed Scopus (66) Google Scholar) to estimate effect sizes of each variant-gene pair accounting for covariates used in single-tissue analyses as well as tissue sharing among individuals. All SNPs within 100 kb of 20,488 genes expressed in either iPSCs or HLCs were used for meta-analysis. The probability of tissue specificity was estimated using m values (Han and Eskin, 2012Han B. Eskin E. Interpreting meta-analyses of genome-wide association studies.PLoS Genet. 2012; 8: e1002555Crossref PubMed Scopus (104) Google Scholar). Briefly, m values indicate the posterior probability that the effect exists in each study or cell type. We considered m values ≥ 0.9 as indicating a particular variant had a molecular effect in each cell type. The most significant HLC eGenes with respect to meta-analysis p values were AMDHD1, FA2H, and ACAA2 (Table S5). We assessed tissue specificity of HLC and iPSC eGenes by selecting SNP-eGene pairs (the most significant single-tissue SNP per gene) (Table S5). Among 2,392 HLC eGenes, we found 267 HLC-specific eGenes (HLC m value ≥ 0.9; iPSC m value < 0.1) and 1,051 eGenes with evidence of tissue specificity in both HLCs and iPSCs (HLC m value ≥ 0.9; iPSC m value ≥ 0.9). Among 3,587 iPSC eGenes, we found 474 iPSC-specific eGenes (HLC m value < 0.1; iPSC m value ≥ 0.9) and 1,279 eGenes with evidence of effects in both HLCs and iPSCs (HLC m value ≥ 0.9; iPSC m value ≥ 0.9). We also indirectly assessed potential HLC and iPSC specificity of GTEx liver eGenes by comparing GTEx liver eGenes to m values estimated from our data. We found 24 GTEx liver eGenes (HLC m value ≥ 0.9; iPSC m value < 0.1) and 10 GTEx liver eGenes (HLC m value < 0.1; iPSC m value ≥ 0.9) that potentially share HLC or iPSC tissue specificity, respectively. Our identified 267 HLC-specific eGenes were significantly enriched for genes with known functions in decreased cholesterol level (MP:0003983; Benjamini-Hochberg FDR = 0.03), decreased sterol level (MP:0012225; FDR = 0.03), and decreased circulating cholesterol level (MP:0005179; FDR = 0.04) (Table S6). Perhaps not surprisingly, these HLC-specific eGenes were also enriched for liver-selective genes as defined by MSigDB (M13283; FDR = 0.04). In order to assess whether the eQTL/ASE analyses could identify new functional variants and genes at GWAS loci, we chose to focus on blood lipid traits (total cholesterol [TC], LDL-C, HDL-C, and triglyceride [TG] levels) because of the major influence of the liver on these traits, as well as the ability to interrogate the functional effects of candidate genes on blood lipid levels in mice. To identify potential functional variants and genes in the HLC, iPSC, and GTEx liver cohorts, we interrogated 7,240 SNPs across 157 loci associated with one or more of the four lipid traits (p < 5 × 10−8) from the Global Lipids Genetics Consortium (GLGC) (Willer et al., 2013Willer C.J. Schmidt E.M. Sengupta S. Peloso G.M. Gustafsson S. Kanoni S. Ganna A. Chen J. Buchkovich M.L. Mora S. et al.Global Lipids Genetics ConsortiumDiscovery and refinement of loci associated with lipid levels.Nat. Genet. 2013; 45: 1274-1283Crossref PubMed Scopus (1875) Google Scholar) for association with the expression of any gene within 1 Mb (eQTL FDR < 5%). Across the 157 loci, there were 32 eGenes in the HLCs, 43 eGenes in the iPSCs, and 30 eGenes in the GTEx livers (Figure 2A; Table S7). In the HLCs, 205 genes at the GLGC loci exhibited significant ASE (FDR < 5%) in at least one individual, potentially identifying additional candidate functional genes that were missed by the eQTL analyses. We were particularly interested in the HLC eGenes, irrespective of whether they were found only in HLCs or were shared by iPSCs (i.e., not exclusive to HLCs), because of their potential to modulate hepatocyte lipid metabolism. We rank-ordered the identified GLGC HLC eGenes by strength of association, judged by eQTL p value (Figure 2B). Among the top GLGC HLC eGenes, we deferred studying SYPL2 and IGF2R; for each of these genes, the best eQTL SNP is in weak LD with an extremely strongly lipid-associated SNP in the SORT1 locus (r2 = 0.01 between rs10857787 and rs12740374 in Europeans) and the LPA locus (r2 = 0.035 between rs3777404 and rs1564348 in Europeans), respectively—so strongly that the SYPL2 and IGF2R SNPs meet the statistical threshold of p < 5 × 10−8 for lipid association even though they are located in different loci than the SORT1 and LPA lead SNPs, simply by being in weak LD with them. We also deferred studying FUT2, since its best eQTL SNP is a coding variant in the gene. Among the other top GLGC HLC eGenes, we sought to identify functional regulatory variants in the loci of CPNE1, ANGPTL3, and FRK, as described below. We note that in eQTL analyses performed only on higher-quality HLCs (the 63 HLC samples with scores of at least three in Table S1, out of the total 86 HLC samples), all of the top GLGC HLC eGenes remained strongly associated with SNPs in their loci (Figure 2B). To assess the hypothesis that lipid and eQTL associations are driven by the same variant, we assessed colocalization of each of eight top GLGC HLC eGenes and the four relevant GLGC traits (Giambartolomei et al., 2014Giambartolomei C. Vukcevic D. Schadt E.E. Franke L. Hingorani A.D. Wallace C. Plagnol V. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics.PLoS Genet. 2014; 10: e1004383Crossref PubMed Scopus (880) Google Scholar). Fifteen trait-eGene pairs exhibited moderate to strong evidence of colocalization (PP.H4.ABF > 0.4; Table S8). In particular, VKORC1 showed strong evidence of colocalization in both HLCs (TG, PP.H4.ABF = 0.97; LDL-C, PP.H4.ABF = 0.52) and GTEx livers (TG, PP.H4.ABF = 0.98; LDL-C, PP.H4.ABF = 0.65), but not in iPSCs (TG, PP.H4.ABF = 0.07; LDL-C, PP.H4.ABF = 0.006). ANGPTL3 colocalized in HLCs with all four GLGC traits (LDL-C, PP.H4.ABF = 0.77; TG, PP.H4.ABF = 0.76; TC, PP.H4.ABF = 0.49; HDL-C, PP.H4.ABF = 0.48). ANGPTL3 did not show evidence of colocalization in iPSCs or GTEx livers (Table S8). We focused first on the locus of CPNE1 (Figure 3A), an HLC eGene that did not independently qualify as an eGene in the GTEx liver cohort (i.e., did not have any associated SNPs with FDR < 5%), although it qualified as an eGene in a cohort of ∼1,000 liver samples (Teslovich et al., 2010Teslovich T.M. Musunuru K. Smith A.V. Edmondson A.C. Stylianou I.M. Koseki M. Pirruccello J.P. Ripatti S. Chasman D.I. Willer C.J. et al.Biological, clinical and population relevance of 95 loci for blood lipids.Nature. 2010; 466: 707-713Crossref PubMed Scopus (2781) Google Scholar). To identify candidate functional variants among all the SNPs in LD within the CPNE1 locus in an unbiased fashion, we performed MPRA experiments to rapidly profile the regulatory activity of 239 variants linked (r2 ≥ 0.5) to the lead SNP at the locus. Duplicate MPRAs were performed, and we rank-ordered SNPs by magnitude of allele-specific regulatory activity as measured by reporter expression (Table S9). The top MPRA SNP in the locus was rs2277862 (Figure 2C). Of note, rs2277862 was also the lead SNP for TC in the locus in the GLGC study, with each alternate allele copy associated with a 1.19 mg/dL change in TC (Teslovich et al., 2010Teslovich T.M. Musunuru K. Smith A.V. Edmondson A.C. Stylianou I.M. Koseki M. Pirruccello J.P. Ripatti S. Chasman D.I. Willer C.J. et al.Biological, clinical and population relevance of 95 loci for blood lipids.Nature. 2010; 466: 707-713Crossref PubMed Scopus (2781) Google Scholar). The locus harbors eGenes in the HLC cohort (CPNE1), the iPSC cohort (CPNE1), and the GTEx liver cohort (ERGIC3); CPNE1 exhibited ASE in HLCs and iPSCs (lowest CPNE1 HLC p = 3 × 10−8; lowest CPNE1 iPSC p = 6 × 10−9). The functions of these genes are poorly understood. We reasoned that rs2277862 might modulate transcription of CPNE1 in both HLCs and undifferentiated hPSCs. To test if rs2277862 regulates CPNE1 and/or ERGIC3 expression, we performed four types of experiments, two based in hPSCs. First, in an hPSC line, HUES 8 (C/C, major/major at rs2277862), we used CRISPR-Cas9 to knock in one minor allele (Figure 3B). Using a single-strand DNA oligonucleotide as a repair template, we obtained only a single recombinant heterozygous (C/T) clone at a frequency of 0.15% (1 out of 672 clones screened), highlighting the inefficiency of the procedure. We observed significantly decreased CPNE1 expression in both undifferentiated knockin hPSCs (down 19%) and differentiated knockin HLCs (down 10%), with non-significant effects on ERGIC3 (Figure 3C). Second, in two hPSC lines, HUES 8 and H7 (T/T, minor/minor at rs2277862), we used multiplexed CRISPR-Cas9 to cleanly and efficiently delete ∼38 bp around the SNP (168/285 clones) (Figure 3B), which was far more efficient than knockin. Homozygous deletion of 38 bp in undifferentiated HUES 8 cells decreased expression of CPNE1 (down 31%) and ERGIC3 (down 20%), whereas homozygous deletion in undifferentiated H7 cells did not affect CPNE1 and decreased expression of ERGIC3 to a lesser degree (down 10%) (Figure 3C). These data are concordant with the data from the heterozygous knockin clone. Third, we used catalytically dead Cas9 (dCas9) with three different guide RNAs (gRNAs) at the site of the SNP to attempt CRISPR interference, or CRISPRi (Qi et al., 2013Qi L.S. Larson M.H. Gilbert L.A. Doudna J.A. Weissman J.S. Arkin A.P. Lim W.A. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression.Cell. 2013; 152: 1173-1183Abstract Full Text Full Text PDF PubMed Scopus (2996) Google Scholar), in HEK293T cells (C/C, major/major at rs2277862). The rationale was that if the variant were truly functional and lay within a transcriptional regulatory element, then the presence of the bulky Cas9 protein at the site should sterically hinder recruitment of transcription factors to the regulatory site and thus interfere with transcriptional regulation. We found that the gRNAs generally resulted in decreased expression of CPNE1, but not ERGIC3 (Figure 3D). Fourth, we noted that, atypically, the noncoding region encompassing rs2277862 is well conserved in mouse (Figures 3A and 3E), and the orthologous nucleotide in mouse also displays naturally occurring variation and has been previously cataloged as rs27324996, with the same two alleles (C and T) as in humans. We used CRISPR-Cas9 in single-cell mouse embryos of the C57BL/6J strain (C/C at rs27324996) to generate a knockin mouse with a minor allele (T) as well as four additional nucleotide changes both to prevent CRISPR-Cas9 re-cleavage of the knockin allele and to “humanize” the sequence (i.e., make a perfect match to the orthologous human sequence) (Figure 3E). There was decreased expression of Cpne1 (down 37%) and Ergic3 (down 34%, albeit not with p < 0.05) in the liver in homozygous knockin (T/T) mice compared to wild-type (C/C) littermates (Figure 3F); cholesterol levels were unchanged, as expected from the small effect in humans (1.19 mg/dL change in cholesterol per allele). These data are highly concordant with the data from the genome-edited hPSC clones with respect to both the direction and magnitude of effects. Taken together, all of these findings support rs2277862-CPNE1 as a functional SNP-gene set. We next focused on the locus of ANGPTL3 (Figure 4A), another HLC eGene that did not independently qualify as an eGene in the GTEx liver cohort, although it qualified as an eGene in a cohort of ∼1000 liver samples (Teslovich et al., 2010Teslovich T.M. Musunuru K. Smith A.V. Edmondson A.C. Stylianou I.M. Koseki M. Pirruccello J.P. Ripatti S. Chasman D.I. Willer C.J. et al.Biological, clinical and population relevance of 95 loci for blood lipids.Nature. 2010; 466: 707-713Crossref PubMed Scopus (2781) Google Scholar). In duplicate MPRA experiments, we profiled the regulatory activity of 210 variants linked (r2 ≥ 0.5) to the lead SNP at the locus (Table S9). The top MPRA SNP in the locus was rs10889356 (Figure 2C). In Europeans, rs10889356 is tightly linked to rs2131925 (r2 = 0.90), the lead SNP in the locus in the GLGC study, with each alternate allele copy of rs2131925 associated with a 4.9 mg/dL change in TG (Teslovich et al., 2010Teslovich T.M. Musunuru K. Smith A.V. Edmondson A.C. Stylianou I.M. Koseki M. Pirruccello J.P. Ripatti S. Chasman D.I. Willer C.J. et al.Biological, clinical and population relevance of 95 loci for blood lipids.Nature. 2010; 466: 707-713Crossref PubMed Scopus (2781) Google Scholar). ANGPTL3, which is generally believed to be the functional gene at this locus, encodes a liver-specific secreted protein that increases blood cholesterol and triglyceride levels in mice (Koishi et al., 2002Koishi R. Ando Y. Ono M. Shimamura M. Yasumo H. Fujiwara T. Horikoshi H. Furukawa H. Angptl3 regulates lipid metabolism in mice.Nat. Genet. 2002; 30: 151-157Crossref PubMed Scopus (325) Google Scholar) and in humans (Musunuru et al., 2010Musunuru K. Pirruccello J.P. Do R. Peloso G.M. Guiducci C. Sougnez C. Garimella K.V. Fisher S. Abreu J. Barry A.J. et al.Exome sequencing, ANGPTL3 mutations, and familial combined hypolipidemia.N. Engl. J. Med. 2010; 363: 2220-2227Crossref PubMed Scopus (503) Google Scholar). ANGPTL3 lies within an intron of DOCK7, a poorly understood non-liver-specific gene. As with the CPNE1/ERGIC3 locus, we used several complementary approaches to interrogate the relationship of rs10889356 with ANGPTL3 and DOCK7. First, in the H7 hPSC line (G/G, major/major at rs10889356), we used CRISPR-Cas9 to knock in the minor allele (Figure 4B). We attempted to use a single-strand DNA oligonucleotide as a repair template but were unsuccessful in obtaining any heterozygous or homozygous knockin clones after screening a large number of clones. We thereafter used a targeting vector with 500-bp homology arms and a puromycin resistance cassette on a transposon that could undergo scarless removal from a TTAA site with piggyBac (Figure 4B). The closest endogenous TTAA site to rs10889356 was 200 bp away. We obtained several clones in which puromycin selection facilitated knockin of the rs10889356 minor allele into both chromosomes (i.e., homozygous knockin). We pooled these clones and introduced piggyBac, which yielded a large number of clones in which scarless excision was achieved. Homozygous knockin (A/A) clones had decreased expression of DOCK7 compared to wild-type clones (down 16%) (Figure 4D). When differentiated into HLCs, homozygous knockin cells had decreased expression of DOCK7 (down 36%) and substantially increased ANGPTL3 expression (up 60%) (Figure 4D). Second, in the same H7 hPSC background (G/G, major/major), we used multiplexed CRISPR-Cas9 to delete 36–39 bp around the SNP (Figure 4C). Homozygous deletion of 36–39 bp in undifferentiated H7 cells decreased expression of DOCK7 (down 8%) (Figure 4E). When differentiated into HLCs, homozygous deleted H7 cells had decreased expression of DOCK7 (down 32%) and increased ANGPTL3 expression (up 67%) (Figure 4E). These data are extremely concordant with the data from the homozygous knockin clones. Third, CRISPRi with three gRNAs in HepG2 cells (G/G, major/major at rs10889356; used because being of hepatic origin they express ANGPTL3) both decreased DOCK7 expression and increased ANGPTL3 expression (Figure 4F). These data are concordant with the data from the genome-edited hPSC clones with respect to the direction of effects for both genes. Taken together, all of these findings support rs10889356-DOCK7 and rs10889356-ANGPTL3 as functional SNP-gene sets. We investigated a third locus with a hig" @default.
• W2604964915 created "2017-04-14" @default.
• W2604964915 creator A5001181025 @default.
• W2604964915 creator A5001583069 @default.
• W2604964915 creator A5006789814 @default.
• W2604964915 creator A5010484062 @default.
• W2604964915 creator A5012278873 @default.
• W2604964915 creator A5016241700 @default.
• W2604964915 creator A5018693048 @default.
• W2604964915 creator A5026897418 @default.
• W2604964915 creator A5030911520 @default.
• W2604964915 creator A5032419354 @default.
• W2604964915 creator A5035376263 @default.
• W2604964915 creator A5040568538 @default.
• W2604964915 creator A5043655881 @default.
• W2604964915 creator A5045216788 @default.
• W2604964915 creator A5053586170 @default.
• W2604964915 creator A5053914271 @default.
• W2604964915 creator A5054758029 @default.
• W2604964915 creator A5055029545 @default.
• W2604964915 creator A5058010200 @default.
• W2604964915 creator A5061926947 @default.
• W2604964915 creator A5065670232 @default.
• W2604964915 creator A5067149045 @default.
• W2604964915 creator A5068633258 @default.
• W2604964915 creator A5072051796 @default.
• W2604964915 creator A5078653034 @default.
• W2604964915 creator A5080121245 @default.
• W2604964915 creator A5083757642 @default.
• W2604964915 creator A5085432309 @default.
• W2604964915 creator A5087046731 @default.
• W2604964915 creator A5088788338 @default.
• W2604964915 date "2017-04-01" @default.
• W2604964915 modified "2023-10-11" @default.
• W2604964915 title "Large, Diverse Population Cohorts of hiPSCs and Derived Hepatocyte-like Cells Reveal Functional Genetic Variation at Blood Lipid-Associated Loci" @default.
• W2604964915 cites W1533942137 @default.
• W2604964915 cites W1978116054 @default.
• W2604964915 cites W1989004527 @default.
• W2604964915 cites W1992436001 @default.
• W2604964915 cites W1999574084 @default.
• W2604964915 cites W2001320379 @default.
• W2604964915 cites W2004445343 @default.
• W2604964915 cites W2013620034 @default.
• W2604964915 cites W2023386301 @default.
• W2604964915 cites W2023839452 @default.
• W2604964915 cites W2048612325 @default.
• W2604964915 cites W2058804806 @default.
• W2604964915 cites W2068427957 @default.
• W2604964915 cites W2073836327 @default.
• W2604964915 cites W2076154138 @default.
• W2604964915 cites W2086493366 @default.
• W2604964915 cites W2093529520 @default.
• W2604964915 cites W2100697281 @default.
• W2604964915 cites W2101749257 @default.
• W2604964915 cites W2102127267 @default.
• W2604964915 cites W2106937178 @default.
• W2604964915 cites W2108234281 @default.
• W2604964915 cites W2115801316 @default.
• W2604964915 cites W2124649657 @default.
• W2604964915 cites W2147215182 @default.
• W2604964915 cites W2152435464 @default.
• W2604964915 cites W2153118028 @default.
• W2604964915 cites W2157752701 @default.
• W2604964915 cites W2160051211 @default.
• W2604964915 cites W2167942713 @default.
• W2604964915 cites W2169456326 @default.
• W2604964915 cites W2179438025 @default.
• W2604964915 cites W2187181884 @default.
• W2604964915 cites W2210094278 @default.
• W2604964915 cites W2221356560 @default.
• W2604964915 cites W2338408645 @default.
• W2604964915 cites W2412481123 @default.
• W2604964915 cites W2413629639 @default.
• W2604964915 cites W2566843361 @default.
• W2604964915 cites W2949145039 @default.
• W2604964915 doi "https://doi.org/10.1016/j.stem.2017.03.017" @default.
• W2604964915 hasPubMedCentralId "https://www.ncbi.nlm.nih.gov/pmc/articles/5476422" @default.
• W2604964915 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/28388432" @default.
• W2604964915 hasPublicationYear "2017" @default.
• W2604964915 type Work @default.
• W2604964915 sameAs 2604964915 @default.
• W2604964915 citedByCount "132" @default.
• W2604964915 countsByYear W26049649152017 @default.
• W2604964915 countsByYear W26049649152018 @default.
• W2604964915 countsByYear W26049649152019 @default.
• W2604964915 countsByYear W26049649152020 @default.
• W2604964915 countsByYear W26049649152021 @default.
• W2604964915 countsByYear W26049649152022 @default.
• W2604964915 countsByYear W26049649152023 @default.
• W2604964915 crossrefType "journal-article" @default.
• W2604964915 hasAuthorship W2604964915A5001181025 @default.
• W2604964915 hasAuthorship W2604964915A5001583069 @default.
• W2604964915 hasAuthorship W2604964915A5006789814 @default.
• W2604964915 hasAuthorship W2604964915A5010484062 @default.
• W2604964915 hasAuthorship W2604964915A5012278873 @default.
• W2604964915 hasAuthorship W2604964915A5016241700 @default.
• W2604964915 hasAuthorship W2604964915A5018693048 @default.
• W2604964915 hasAuthorship W2604964915A5026897418 @default.

• next